Matrix Metalloproteinases (Biology of Extracellular Matrix)

Matrix Metalloproteinases

Biology of Extracellular Matrix Series Editor R O B E R T P. M E C H A M

Robert P. Mecham: REGULATION OF MATRIX ACCUMULATION Thomas N. Wight and Robert P. Mecham: BIOLOGY OF PROTEOGLYCANS Richard Mayne and Robert E. Burgeson: STRUCTURE AND FUNCTION OF COLLAGEN TYPES Deane R. Mosher: FIBRONECTIN W. Steven Adair and Robert P. Mecham: ORGANIZATION AND ASSEMBLY OF PLANT AND ANIMAL EXTRACELLULAR MATRIX Linda J. Sandell and Charles D. Boyd: EXTRACELLULAR MATRIX GENES John A. McDonald and Robert P. Mecham: RECEPTORS FOR EXTRACELLULAR MATRIX David D. Roberts and Robert P. Mecham: CELL SURFACE AND EXTRACELLULAR GLYCOCONJUGATES Peter D. Yurchenco, David E. Birk, and Robert P. Mecham: EXTRACELLULAR MATRIX ASSEMBLY AND STRUCTURE David A. Cheresh and Robert P. Mecham: INTEGRINS: MOLECULAR AND BIOLOGICAL RESPONSES TO THE EXTRACELLULAR MATRIX

MATRIX METALL O PROTEINAS ES

E d i t e d by

WILLIAM C. PARKS Departments of Medicine and Cell Biology and Physiology Washington University School of Medicine St. Louis, Missouri

ROBERT P. MECHAM Department of Cell Biology and Physiology Washington University School of Medicine St. Louis, Missouri

ACADEMIC PRESS San Diego London Boston New York Sydney Tokyo Toronto

Front cover photographs: (Top) In situ hybridization showing expression of collagenase-1 (MMP-1) by adventitial fibroblasts in a developing pulmonary artery from a 210-day fetal calf. Photomicrograph provided by William C. Parks. (Center) As detected by in situ hybridization, matrilysin (MMP-7) is prominently expressed by Paneth cells in the crypts of mouse ileum. Photomicrograph from Wilson et al., 1995, used with permission. (Bottom) At the point of epidermal separation in blister formation, gelatinase-B (MMP-9) is produced and released by eosinophils. Photomicrograph provided by Mona St~ihle-B~ickdahl. Wilson, C. L., Heppner, K. J., Rudolph, L. A., and Matrisian, L. M. (1995). The metalloproteinase matrilysin is preferentially expressed by epithelial cells in a tissue-restricted pattern in the mouse. Mol. Biol. Cell 6, 8 5 1 - 869.

This book is printed on acid-free paper.

Copyright 9 1998 by ACADEMIC PRESS All Rights Reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the Publisher. The appearance of the code at the bottom of the first page of a chapter in this book indicates the Publisher's consent that copies of the chapter may be made for personal or internal use of specific clients. This consent is given on the condition, however, that the copier pay the stated per copy fee through the Copyright Clearance Center, Inc. (222 Rosewood Drive, Danvers, Massachusetts 01923), for copying beyond that permitted by Sections 107 or 108 of the U.S. Copyright Law. This consent does not extend to other kinds of copying, such as copying for general distribution, for advertising or promotional purposes, for creating new collective works, or for resale. Copy fees for pre-1998 chapters are as shown on the title pages. If no fee code appears on the title page, the copy fee is the same as for current chapters. 0887-3224/98 $25.00 Academic Press a division of Harcourt Brace & Company 525 B Street, Suite 1900, San Diego, California 92101-4495, USA http ://ww w. apnet.com Academic Press Limited 24-28 Oval Road, London NW1 7DX, UK http://www.hbuk.co.uk/ap/ International Standard Book Number: 0-12-545090-7

PRINTED IN THE UNITED STATES OF AMERICA 98 99 00 01 02 03 BB 9 8 7 6

5

4

3 2

1

Contents

.....................................................................................

ix

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .....................................

xi

CONTRIBUTORS

PREFACE

The Matrix Metalloproteinase Family J. FREDERICK WOESSNER, JR. I. I n t r o d u c t i o n .................................................................................... II. T h e P o s i t i o n o f M M P s w i t h i n t h e C l a s s of M e t a l l o p r o t e a s e s ........................................................................ III. T h e M a t r i x i n S u b f a m i l y ................................................................ IV. T h e D o m a i n S t r u c t u r e o f t h e M a t r i x i n s ...................................... V. T h e B i o l o g i c a l R o l e of t h e M a t r i x i n s ............................................ VI. W h a t t h e F u t u r e M i g h t H o l d ........................................................ References ......................................................................................

1 3 4 7 9 12 13

Interstitial Collagenases JOHN J. JEFFERY I. I n t r o d u c t i o n .................................................................................... II. M a t r i x M e t a l l o p r o t e i n a s e - l : T h e O r i g i n a l Interstitial Collagenase ................................................................. III. " N o n t r a d i t i o n a l " I n t e r s t i t i a l C o l l a g e n a s e s .................................. IV. S u m m a r y a n d F u t u r e P e r s p e c t i v e s .............................................. References ......................................................................................

15 16 35 37 38

Stromelysins 1 and 2 HIDEAKI NAGASE I. II. III. IV. V. VI. VII.

Introduction .................................................................................... Structure and Substrate Specificity ............................................. Activation Mechanisms ................................................................. Inhibitors ........................................................................................ R e g u l a t i o n of G e n e E x p r e s s i o n ..................................................... Biological and Pathological Roles ................................................. Conclusions and Future Prospects ............................................... References ......................................................................................

43 44 51 54 56 60 67 68

vi

CONTENTS

72-kDa Gelatinase (Gelatinase A): Structure, Activation, Regulation, and Substrate Specificity ANITA E. YU, ANNE N. MURPHY, and WILLIAM G. STETLER-STEVENSON I. II. III. IV. V. VI. VII.

Introduction .................................................................................... Matrix Metalloproteinase Family ................................................. M u l t i l e v e l R e g u l a t i o n of G e l a t i n a s e A ......................................... S u b s t r a t e Specificity ...................................................................... Role of G e l a t i n a s e A in C e l l u l a r P r o c e s s e s ................................. Role of G e l a t i n a s e A in P l a t e l e t A g g r e g a t i o n .............................. Perspectives .................................................................................... References ......................................................................................

85 86 88 97 99 105 105 106

Gelatinase B: Structure, Regulation, and Function THIENNU H. VU and ZENA WERB I. II. III. IV. V. VI. VII.

Introduction .................................................................................... Gene Structure ............................................................................... Protein Structure and Activation ................................................. Substrates and Inhibitors ............................................................. R e g u l a t i o n of E x p r e s s i o n .............................................................. Function .......................................................................................... Conclusion and Future Directions ................................................ References ......................................................................................

115 116 119 122 124 128 136 137

Matrilysin CAROLE L. WILSON and LYNN M. MATRISIAN I. II. III. IV. V. VI. VII.

Introduction .................................................................................... A m i n o Acid S e q u e n c e a n d G e n e O r g a n i z a t i o n ............................ R e g u l a t i o n of E x p r e s s i o n .............................................................. Protein Structure and Proteolytic Activities ............................... E x p r e s s i o n in C a n c e r a n d O t h e r D i s e a s e s .................................. Role in N o r m a l T i s s u e R e m o d e l i n g a n d H o m e o s t a s i s ................ Summary and Concluding Remarks ............................................ References ......................................................................................

149 149 152 157 164 169 174 177

Macrophage Elastase (MMP-12) STEVEN D. SHAPIRO and ROBERT M. SENIOR I. II. III. IV.

Introduction .................................................................................... T h e G e n e for M a c r o p h a g e E l a s t a s e a n d I t s E x p r e s s i o n ............. Activation and Processing ............................................................. Protein Preparation .......................................................................

185 187 187 188

V. VI. VII. VIII.

CONTENTS

vii

Catalytic Features ......................................................................... Inhibition ........................................................................................ Biological A s p e c t s .......................................................................... Summary ........................................................................................ References ......................................................................................

188 190 190 195 196

Membrane-Type Matrix Metalloproteinases and Cell SurfaceAssociated Activation Cascades for Matrix Metalloproteinases VERA KNAUPER and GILLIAN MURPHY I. II. III. IV. V. VI. VII. VIII. IX.

Introduction .................................................................................... P o t e n t i a l A c t i v a t i o n R o u t e s for M a t r i x M e t a l l o p r o t e i n a s e s ....... S t r u c t u r e of M e m b r a n e - T y p e M a t r i x M e t a l l o p r o t e i n a s e s .......... C e l l u l a r S o u r c e s of M T - M M P s a n d R e g u l a t i o n of E x p r e s s i o n .................................................................................. A c t i v a t i o n of M T - M M P s ................................................................ A c t i v a t i o n of P r o g e l a t i n a s e A b y M T I - M M P and MT2-MMP ............................................................................... A c t i v a t i o n of P r o c o l l a g e n a s e - 3 b y M T I - M M P ............................. S u b s t r a t e Specificity of M T - M M P s ............................................... Summary ........................................................................................ References ......................................................................................

199 200 202 204 206 207 210 211 212 214

Substrate Specificity and Mechanisms of Substrate Recognition of the Matrix Metalloproteinases VERA IMPER and HAROLD E. VAN WART I. II. III. IV.

Introduction .................................................................................... P e p t i d e S e q u e n c e Specificity ......................................................... P r o t e i n S u b s t r a t e Specificity ........................................................ M e c h a n i s m s of S u b s t r a t e R e c o g n i t i o n ......................................... References ......................................................................................

219 221 228 233 238

Synthetic Inhibitors of Matrix Metalloproteinases PETER D. BROWN I. II. III. IV.

Introduction .................................................................................... D e s i g n of S y n t h e t i c I n h i b i t o r s ...................................................... Role of M a t r i x M e t a l l o p r o t e i n a s e s in C a n c e r .............................. E f f e c t s of M a t r i x M e t a l l o p r o t e i n a s e I n h i b i t o r s in Cancer Models ................................................................................ V. C l i n i c a l T r i a l s ................................................................................ VI. P r o t e i n a s e I n h i b i t i o n as T h e r a p y ................................................. References ......................................................................................

243 244 247 249 253 255 256

viii

CONTENTS

Matrix Metalloproteinases in Tissue Repair WILLIAM C. PARKS, BARRY D. SUDBECK, GLENN R. DOYLE, and ULPU K. SAARIAHLO-KERE I. I n t r o d u c t i o n : S t u d y i n g t h e R e g u l a t i o n a n d F u n c t i o n of M e t a l l o p r o t e i n a s e s .................................................................... II. C o l l a g e n a s e - 1 in T i s s u e R e p a i r .................................................... III. T h e Role of C o l l a g e n a s e - 1 in R e e p i t h e l i a l i z a t i o n ........................ IV. E x p r e s s i o n of O t h e r M e t a l l o p r o t e i n a s e s in W o u n d R e p a i r ........ V. R e g u l a t i o n of C o l l a g e n a s e - 1 in W o u n d F i b r o b l a s t s .................... VI. M e t a l l o p r o t e i n a s e s in C h r o n i c W o u n d s ....................................... VII. S u m m a r y ........................................................................................ References ......................................................................................

263 266 277 283 286 288 289 290

Regulation of Matrix Metalloproteinase Gene Expression M. ELIZABETH FINI, JEFFERY R. COOK, ROYCE MOHAN, and CONSTANCE E. BRINCKERHOFF I. II. III. IV. V. VI.

Introduction .................................................................................... Collagenase .................................................................................... Stromelysins ................................................................................... Gelatinase B ................................................................................... Gelatinase A ................................................................................... Summary and Future Perspectives .............................................. References ......................................................................................

INDEX

........................................................................................................

300 302 321 327 334 338 339

357

Contributors

Numbers in parentheses indicate the pages on which the authors" contributions begin.

CONSTANCE E. BRINCKERHOFF (299), Departments of Medicine and Biochemistry, Dartmouth Medical School, Hanover, New Hampshire 03755 PETER D. BROWN (243), Department of Clinical Research, British Biotech Pharmaceutical Ltd., Oxford OX4 5LY, United Kingdom JEFFERY R. COOK (299), Vision Research Laboratories of the New England Medical Center and Departments of Ophthalmology and Anatomy and Cellular Biology, Tufts University School of Medicine, Boston, Massachusetts 02111 GLENN R. DOYLE (263), Departments of Medicine and Cell Biology and Physiology, Washington University School of Medicine, St. Louis, Missouri 63110 M. ELIZABETH FINI (299), Vision Research Laboratories of the New England Medical Center and Departments of Ophthalmology and Anatomy and Cellular Biology, Tufts University School of Medicine, Boston, Massachusetts 02111 VERA IMPER (219), Faculty of Pharmacy and Biochemistry, University of Zagreb, 10000 Zagreb, Croatia JOHN J. JEFFERY (15), Department of Biochemistry and Molecular Biology, Albany Medical College, Albany, New York 12208 VERA KNAUPER (199), Strangeways Research Laboratory, Cambridge, and School of Biological Sciences, University of East Anglia, Norwich NR4 7TJ, United Kingdom LYNN M. MATRISIAN (149), Department of Cell Biology, Vanderbilt University School of Medicine, Nashville, Tennessee 37232 ROYCE MOHAN (299), Vision Research Laboratories of the New England Medical Center and Departments of Ophthalmology and Anatomy and Cellular Biology, Tufts University School of Medicine, Boston, Massachusetts 02111 ANNE N. MURPHY (85), Department of Biochemistry, George Washington University Medical School, Washington, DC 20037 ix

x

CONTRIBUTORS

GILLIAN MURPHY (199), Strangeways Research Laboratory, Cambridge, and School of Biological Sciences, University of East Anglia, Norwich NR4 7TJ, United Kingdom HIDEAKI NAGASE (43), Department of Biochemistry and Molecular Biology, University of Kansas Medical Center, Kansas City, Kansas 66160 WILLIAM C. PARKS (263), Departments of Medicine and Cell Biology and Physiology, Washington University School of Medicine, St. Louis, Missouri 63110 ULPU K. SAARIAHLO-KERE (263), Department of Dermatology, University of Helsinki, Helsinki, Finland ROBERT M. SENIOR (185), Department of Medicine, Washington University School of Medicine at Barnes-Jewish Hospital, St. Louis, Missouri 63110 STEVEN D. SHAPIRO (185), Departments of Medicine and Cell Biology, Washington University School of Medicine, St. Louis, Missouri 63110 WILLIAM G. STETLER-STEVENSON (85), Extracellular Matrix Pathology Section, Laboratory of Pathology, National Cancer Institute, National Institutes of Health, Bethesda, Maryland 20892 BARRY D. SUDBECK (263), Departments of Medicine and Cell Biology and Physiology, Washington University School of Medicine, St. Louis, Missouri 63110 HAROLD E. VAN WART (219), Inflammatory Diseases Unit, $3-1, Roche Bioscience, Palo Alto, California 94304 THIENNU H. VU (115), Department of Medicine, Division of Pulmonary and Critical Care Medicine, University of California, San Francisco, California 94143 ZENA WERB (115), Department of Anatomy, University of California, San Francisco, California 94143 CAROLE L. WILSON (149), Dermatology Division, Washington University School of Medicine, St. Louis, Missouri 63110 J. FREDERICK WOESSNER, JR. (1), Department of Biochemistry and Molecular Biology, University of Miami School of Medicine, Miami, Florida 33101 ANITA E. YU (85), Extracellular Matrix Pathology Section, Laboratory of Pathology, National Cancer Institute, National Institutes of Health, Bethesda, Maryland 20892

Preface

The spatially and temporally precise removal and remodeling of connective tissue are critical to several developmental, homeostatic, and reparative processes. Matrix turnover requires the activity of many different endopeptidases acting on a variety of compositionally distinct proteins, and consequently, it is not surprising that over the past decades numerous proteinases of distinct gene families have been characterized and implicated as serving a catalytic function during tissue remodeling. For example, the large and physiologically important serine proteinase family, which includes leukocyte elastase, plasminogen, and its activators, among many other enzymes, mediates a variety of activities, from clot dissolution to tissue destruction. The matrix metalloproteinases (MMPs), which compose the matrixin subfamily of the large metalloproteinase family (see Chapter 1) and which are the focus of this book, have a specialized function in turnover of the extracellular matrix. As suggested throughout this volume, however, the activity of these enzymes may also be involved in other functions, such as protein processing and activation. In recent years, MMPs have gained considerable attention in many studies on normal tissue events, inflammation, and disease processes, and this enhanced interest is probably due to several factors. For the most part, MMPs are produced or activated when needed, and expression of these enzymes provides a reliable indicator of ongoing tissue remodeling. Thus, for investigators, these enzymes are models of gene products that are accurately regulated and precisely targeted to specific extracellular substrates by a wide variety of cells during numerous normal tissue processes, such as wound healing, bone resorption, and morphogenesis. In contrast, exuberant production of MMPs is a hallm a r k of many destructive diseases, such as arthritis and chronic ulcerations, and of many disease-related processes, such as inflammation, metastasis, and angiogenesis, and aberrant regulation of MMP production is thought to be a primary mechanism contributing to disease progression and injury. As such, understanding how MMP expression and activation are controlled and identifying which proteinases are produced by which cells under defined conditions will have a great impact on our understanding of normal biology and disease pathogenesis. xi

xii

PREFACE

The metalloproteinase field has been expanded by a number of important and interesting discoveries. For example, activation of latent tumor necrosis factor-c~ (TNF-a), selectin shedding, cleavage of collagen N-propeptide, and processing of laminin-5 chains are all conferred by metalloproteinases but by enzymes that are members of subfamilies distinct from MMPs. In this volume, we focus on the matrix metalloproteinases, which are believed to act in the extracellular space and to have a specialized role in the degradation of connective tissue proteins. Although the ADAMs, such as TNF convertase and others, constitute an emerging and very interesting field in themselves, we have not included these in this volume, as they have not been implicated as having a direct role in matrix turnover. Stromelysin-3 (MMP-11) is also not included, because it apparently has no significant activity for matrix components. This book begins with a chapter by J. Frederick Woessner, Jr., who outlines the distinct and specific features of the matrixin subfamily of metalloproteinases. The book then provides up-to-date reviews on the structure, biochemistry, function, and molecular biology of each of the principal MMPs written by investigators who are actively studying these enzymes (Chapters 2-8). Knowledge of the crystal structure and distinct substrate specificity of different MMPs, most of which has been gained over the past few years, has provided needed information for cogent design of synthetic inhibitors of metalloproteinase activity, and recent work in this field is discussed in chapters by Harold Van Wart and Peter Brown (Chapters 9 and 10). This book also includes two chapters on regulation of MMP expression at the level of the gene and by influences arising in the tissue environment (Chapters 11 and 12). We thank the many authors for their thoughtful and timely contributions, and we hope this book proves helpful for the many investigators and students who have an interest in metalloproteinase biology. William C. Parks Robert P. Mecham

The Matrix Metalloproteinase Family J. Frederick Woessner, Jr. Department of Biochemistry and Molecular Biology, University of Miami School of Medicine, Miami, Florida 33101

I. Introduction II. The Position of MMPs within the Class of Metalloproteases III. The Matrixin Subfamily A. Members of the Subfamily B. Evolutionary Relationships among the MMPs IV. The Domain Structure of the Matrixins V. The Biological Role of the Matrixins A. Do the MMPs Play a Major Role in Degrading the Extracellular Matrix? B. What Are the True Substrates for Each MMP? C. Why So Many MMPs? VI. What the Future Might Hold References

I.

INTRODUCTION

Seldom has a field had such a clearly defined beginning as did the field of matrix metalloproteinases (MMPs). The first report of an enzyme from vertebrate (as opposed to bacterial) sources that was capable of attacking the triple helix of native type I collagen was published in 1962 by Jerome Gross and Charles Lapiere. In this first report they demonstrated that the enzyme activity, secreted by cultured tissue fragments of tail fin skin from resorbing tadpole tails in metamorphosis, was a true collagenase acting on collagen at 27~ at neutral pH. They also found activity produced by cultured chick embryo skin, postpartum rat uterus, and mouse and rat bone. Activity was not found in tissue extracts. In retrospect this is now known to be due both to the ability of MMPs to bind to the extracellular matrix and to the stimulation of high enzyme secretion that occurs when tissues are excised and cultured. Also in retrospect we see that at least collagenase 1 (MMP-1) and 3 (MMP-13) were present and possibly collagenase 4 (MMP-18) in these first experiments. Within a short period of time these studies had been broadened by the work of many groups to include various Matrix Metalloproteinases

1

Copyright 9 1998 by Academic Press. All rights of reproduction in any form reserved.

2

J. FREDERICK WOESSNER, JR.

h u m a n sources including n e u t r o p h i l s (MMP-8) a n d to show t h a t collag e n a s e cut the triple helix at a point o n e - q u a r t e r of t h e d i s t a n c e in from t h e C - t e r m i n a l end a n d t h a t t h e activity w a s m e t a l d e p e n d e n t . Since Gross a n d L a p i e r e ' s 1962 report, t h e r e h a s been a t r e m e n d o u s efflorescence of this field; in p r e p a r i n g for a more detailed overview I h a v e compiled a b i b l i o g r a p h y of 8500 i t e m s , w i t h t h e c u r r e n t r a t e of publication a p p r o a c h i n g 1000 i t e m s per year. The r e a s o n for this interest is pointed out m o s t clearly by Table I, which shows some of t h e points of i n v o l v e m e n t of t h e M M P s in n o r m a l biologic a n d pathologic processes. C a n c e r a n d its m e t a s t a t i c s p r e a d h a v e been of t h e g r e a t e s t c u r r e n t w i d e s p r e a d i n t e r e s t . Not only are a g r e a t m a n y different M M P s involved in the processes of carcinogenesis a n d its a c c o m p a n y i n g rem o d e l i n g or d e s t r u c t i o n of the e x t r a c e l l u l a r m a t r i x , b u t t h e r e is also

TABLE I NORMALANDPATHOLOGICALPROCESSESIN WHICHMMPs ARE IMPLICATED Normal Development

Blastocyst implantation Embryonic development Nerve growth Growth plate cartilage removal Skeletal, bone growth Nerve outgrowth Enamel maturation Primary tooth resorption Reproduction

Endometrial cycling Graafian follicle rupture Luteolysis Cervical dilatation Postpartum uterine involution Mammary gland morphogenesis Mammary gland involution Rupture of fetal membranes Maintenance

Remodeling of bone Hair follicle cycle Wound healing Angiogenesis Apoptosis Nerve regeneration Macrophage function Neutrophil function

Pathological Tissue Destruction

Rheumatoid arthritis Osteoarthritis Cancer invasion Cancer metastasis Decubitus ulcer Gastric ulcer Corneal ulceration Periodontal disease Fibrotic Diseases

Liver cirrhosis Fibrotic lung disease Otosclerosis Atherosclerosis Multiple sclerosis

Weakening of Matrix

Dilated cardiomyopathy Epidermolysis bullosa Aortic aneurysm

THE MATRIX METALLOPROTEINASE FAMILY

3

hope of countering this destruction by various approaches such as systemic administration of inhibitors of MMPs, gene therapy to knock out enzymes, or overexpression of natural inhibitors such as the TIMPs (tissue inhibitors of metalloproteinases). II.

T H E POSITION OF THE M M P s WITHIN THE CLASS OF METALLOPROTEASES

The proteases comprise both exopeptidases and endopeptidases (proteinases). These hydrolytic enzymes can be divided into four classes based on the catalytic group at their active center: serine/threonine, cysteine, aspartic, and metallo. It is the class of metalloproteases that concerns us here; most of its members depend on zinc for their catalytic action. A new compendium of proteases (Barrett et al., 1998) describes 200 metalloproteases, of which only 14 are MMPs. Therefore, it is important to u n d e r s t a n d the relationship of the MMPs to this larger class. A useful, but still provisional, guide has been provided by Rawlings and Barrett (1995), who divide the class into clans (based on similarity of protein fold) and families (based on evolutionary relationships). Currently the metallo class comprises eight clans and some 40 families (Barrett et al., 1998). For example, enzymes in clan MA have the zinc-binding motif HEXXH and the third zinc ligand is Glu. Thermolysin provides the X-ray structure for this clan and 5 families are typified by thermolysin, mycolysin, neprilysin, m e m b r a n e alanyl aminopeptidase, and peptidyl-dipeptidase A (Rawlings and Barrett, 1995); there are currently 32 species of enzymes in this clan (Barrett et al., 1998). This HEXXH motif is widespread and occurs in several additional clans. The clan containing the MMP family is clan MB in which the third zinc ligand is not Glu but r a t h e r a third His residue in the consensus sequence HEXXHXXGXXH. The two major families in this clan are M12, with its subfamilies astacin and reprolysin, and M10, with its subfamilies serralysin and matrixin (MMP). These two families contain more t h a n 60 members and there are two further families of one member each (autolysin and snapalysin; Barrett et al., 1998). The X-ray structure of this clan is now known for a number of species in the four subfamilies. Bode et al. (1994) have given the name m e t z i n c i n s to this group because all contain a conserved Met residue to the carboxy side of the zinc site, which produces a t u r n in the protein chain t h a t provides the base of the active center binding pocket. However, the sequence similarities are not otherwise very close between families M10 and M12. The similarities of the binding pocket have the important consequence that hydroxamate inhibitors designed to block the action of MMPs are frequently found to effectively block members of the M10

4

J. FREDERICK WOESSNER, JR.

family. A classic example is the inhibition of tumor necrosis factor (TNF-c~) convertase, an enzyme now known to belong to the ADAM family within the reprolysin subfamily (ADAM-17; Black et al., 1997). I suspect that the "aggrecanase" activity important in cartilage degradation may also prove to be such an example. It can be seen that less than half of the metallo class is accounted for by the first two clans. Further clans include enzymes with the HEXXH motif and a different third ligand for zinc, the motif HXXEH (pitrilysin), and a variety of additional three-ligand arrangements involving various combinations of three residues selected from His, Asp, and Glu, as well as some cases in which there are two zinc atoms and some in which the ligands remain unknown. Full details may be found in Barrett et al. (1998). The scheme is provisional in that we do not yet know the three-dimensional structure of many of the enzymes, and the ligands that bind zinc have not been confirmed in many cases. Note that there is no uniform practice in this field for the naming of families, so the widely used terms MMP family or matrixin family actually describe a subfamily. An earlier set of criteria (Woessner, 1994) for assignment of a new enzyme to the matrixin family included the display of proteolytic activity, function outside the cell, possession of conserved sequence around cysteine in the propeptide (PRCGxPD) and a zinc-binding consensus sequence of HEXGHXXGXXHS/T. However, members of family 10 also meet all of these criteria except the last. Today, a better criterion would be a cDNA sequence that is sufficiently close to that of collagenase to permit assignment to the matrixin subfamily. Many synthetic inhibitors of MMPs, particularly in the hydroxamate series, inhibit members of family 10 equally well as noted earlier. However, TIMP-1 does not appear to block any members of that family, but does appear to inhibit all MMPs. This provides a second criterion.

III.

A.

THE MATRIXIN SUBFAMILY

Members of the S u b f a m i l y

At the time of this writing (late 1997), 17 enzymes have received MMP numbers. The numbers are not being assigned by an official governing body, so some confusion arises as individual authors assign numbers they believe come next (e.g., Cossins et al., 1996). It was suggested at the 1997 Gordon conference on MMPs that I make the assignments in the future. This plan is feasible only if the new enzymes are discovered by those already working in the field. The MMP numbers are often convenient to use as a shorthand when speaking or writing,

THE MATRIX METALLOPROTEINASE FAMILY

5

but their utility is somewhat diminished by the large number of enzyme species t h a t are coming to light. Table II indicates the most commonly used names, including those recommended by the International Union of Biochemistry and Molecular Biology (through MMP-12). Three enzymes reported earlier (MMP-4, -5, -6) were later found to correspond to known enzymes, so these three numbers have been discontinued and remain vacant. The enzymes are placed into arbitrary groups that originally arose from considerations of the substrates cleaved. This is not a very sound basis, because we have very little information about the n a t u r a l substrates of any of these enzymes (see later discussion). Stromelysin 3, while having some substrates in common with the other two stromelysins, is very distantly related and is activated through a furin site and only indirectly through disruption of the cysteine switch. The membrane-type MMPs have been grouped on the basis of possession of a t r a n s m e m b r a n e domain. Table II also notes a few of the many synonyms for these enzymes and indicates additional features of some MMPs.

B. Evolutionary Relationships among the MMPs A dendrogram showing the relationships among the MMPs known from h u m a n s is presented in Fig. 1. The figure is limited to h u m a n enzymes because the more t h a n 65 known sequences from all species produce a tree that is difficult to see for the forest of entries. Only the approximately 170 residues of the catalytic domain for each enzyme have been used in deriving this tree. The earliest branches include MMP-19 and the four membrane-type MMPs. These MT-MMPs form a tight cluster except for MT4-MMP, which is somewhat earlier than the others. A metalloproteinase of the h u m a n ovary has been reported as a gene sequence (Acc. No. D83647), but nothing is known of its properties. As alluded to earlier, stromelysin 3 arose quite early relative to the other two stromelysins. Stromelysin 3 and the four MT-MMPs have in common the RXKR/RRKR furin cleavage site, suggesting that they may be activated while still in the cell, whereas the remaining enzymes require proteolytic cleavage of their propeptide after the zymogen leaves the cell. It is interesting to speculate t h a t simpler organisms regulated their matrix largely by contact, through cell surface MMPs; then as more extensive and complex extracellular matrices evolved, it became advantageous to secrete the MMPs for action at a distance from the cell. The last 10 enzymes form a cluster of the more modern and more commonly known MMPs. The two gelatinases are first in this group; they contain additional fibronectin-like domains. Matrilysin, the small-

TABLE II MEMBERS OF THE MATRIXINFAMILY Group name Collagenase Collagenase Collagenase Collagenase Collagenase Gelatinase

1 2 3 4

Gelatinase A Gelatinase B

MMP number

EC number

Mr latent/active

MMP-1 MMP-8 MMP-13 MMP-18

EC 3.4.24.7 EC 3.4.24.34

MMP-2 MMP-9

EC 3.4.24.24 EC 3.4.24.25

72,000 66,000 92,000 84,000

Type IV collagenase Type V collagenase

MMP-3 MMP-10 MMP-11

EC 3.4.24.17 EC 3.4.24.22 EC 3.4.24.

57,000 45,000 54,000 44,000 64,000 46,000

Transin Transin-2 RXKR furin cleavage

66,000 72,000 64,000 57,000

54,000 60,000 53,000 53,000

Transmembrane domain and RRKR furin cleavage site

28,000 54,000 54,000 54,000

19,000 22,000 45,000 22,000

52,000 85,000 52,000 53,000

42,000 64,000 42,000 42,000

Notes

Interstitial collagenase Neutrophil collagenase Rodent interstitial collagenase Xenopus

Stromelysin

Stromelysin 1 Stromelysin 2 Stromelysin 31 Membrane-type MT1-MMP MT2-MMP MT3-MMP MT4-MMP Others

Matrilysin Metalloelastase (No trivial name) 2 Enamelysin 3 Nonmammalian X e n o p u s XMMP 4 Envelysin 5 Soybean MMP 6

MMP-14 MMP-15 MMP-16 MMP-17 MMP-7 MMP-12 MMP- 19 MMP-20

EC 3.4.25.33 EC 3.4.24.65

70,000 53,000 63,000 48,000 ? 19,000

Lacks hemopexin

Macrophage elastase

Cys in catalytic domain Sea urchin Protein sequencing

Note: The values of Mr, except for MMP-8, are based on cDNA sequence; glycosylation may increase these values. Values for the active forms of MT-MMPs assume cleavage at the furin site. Names in bold are those recommended by the IUBMB. Certain of these enzymes do not receive further attention in the individual chapters; reference to these is as follows: 1 Basset et al., 1990; 2 Cossins et al., 1996; Pendas et al., 1997; 3 Bartlett et al., 1996; 4 Yang et al., 1997; 5 Lepage and Gache, 1990; 6 McGeehan et al., 1992.

7

THE MATRIX METALLOPROTEINASE FAMILY 50 .I

25

0

I

I

PAM Score MT4-MMP MMP-19

i

Ovary MMP MT3-MMP

......

MT2-MMP

.......

MT1-MMP

m

Stromelysin 3

f ....

Gelatinase A

.......

Gelatinase B

'-

Matrilysin

" Enamelysin Neutrophil collagenase -

Interstitial coUagenase Macrophage elastase

.,

Collagenase3 Stromelysin 2

I

Stromelysin 1

FIG. 1. Dendrogram illustrating the evolutionary relationships among the 17 known matrix metalloproteinases found to date in humans. The sequences were aligned by using the P I L E U P program for the catalytic domain only (exclusive of fibronectin repeats). The tree was generated by the KITSCH algorithm. (Figure kindly provided by Dr. Neil Rawlings.)

est MMP, appears next; the absence of a hemopexin-like domain in this enzyme is probably due to a deletion, rather than to an evolutionary origin prior to the addition of hemopexin. Collagenases 1 and 2 are fairly close, followed shortly by macrophage elastase; stromelysins 1 and 2 come last. The last 7 enzymes are closely related in their domain structure, although macrophage elastase loses its hemopexin domain upon activation. IV.

T H E DOMAIN STRUCTURE OF THE MATRIXINS

The matrixins form an interesting group of enzymes in that there is a central catalytic domain to which have been added a variety of additional domains or short inserts. Matrilysin represents the "minimal" e n z y m e - - i t consists of a signal peptide, a propeptide, and the catalytic domain. No one enzyme has all of the possible building blocks. If one examines the MMPs starting from the N terminus, the following features are seen:

8

J. FREDERICK WOESSNER, JR.

Signal Peptide This is typically a stretch of 17-20 residues, rich in hydrophobic amino acids, that serves as a signal for secretion into the endoplasmic reticulum for eventual export from the cell. All of the MMPs except for MMP-17 (Puente et al., 1996) possess a signal peptide. Propeptide This region contains about 80 amino acids, typically with an N-terminal hydrophobic residue. There is a highly conserved PRCXXPD sequence near the C-terminal end of this segment; this provides the cysteine residue that makes contact with the catalytic zinc atom and maintains the enzyme in its zymogen form. This cysteine is found in all MMPs, including those that have the furin-cleavage site. Furin-Cleavage Site Insert This stretch of about nine residues includes the consensus sequence of RXKR/RRKR that leads to intracellular cleavage by furin. MMP-11, -14,-15, -16, a n d - 1 7 possess this sequence. In the remaining enzymes, a cleavage by external proteases occurs in the middle of the propeptide, partially exposing the zinc and leading to autolytic cleavage of the remainder of the propeptide. The exact site of final cleavage may vary within a given enzyme leading to different degrees of activation, particularly in the case of MMP-1. XMMP contains a much longer insert of 37 residues (similar in sequence to vitronectin) that ends in the RRKR motif (Yang et al., 1997). Catalytic Domain This domain typically contains about 160-170 residues, including sites for the binding of calcium ions and the structural zinc atom. The 50-54 residues at the C-terminal end of the catalytic domain include the site of binding of the catalytic zinc. This involves the highly conserved HEXGHXXGXXHS/T sequence mentioned earlier. MMP-17, however, has Val in place ofSer (Puente et al., 1996). The zinc-binding region is somewhat independent of the remainder of the catalytic domain because various insertions can occur between these two portions. Fibronectin-Like Repeats There are three repeats of the fibronectin type II domain in MMP-2 and MMP-9, inserted in the catalytic domain just ahead of the 50-residue zinc-binding region. These specialized structures aid the binding of enzyme to gelatin substrates. Hinge Region The catalytic domain is connected to the following hemopexin domain by a linker region usually referred to as the hinge region. It ranges in length from 0 to 75 residues. The longest hinge is found in MMP-9 and shows considerable homology to type V collagen in that it is rich in proline. MMP-7, having no hemopexin domain, has no need for a hinge and XMMP also lacks this insert. A typical hinge contains about 16 residues including a number of proline residues. MMP-19 has a highly acidic region DEEEEETE within its linker (Cossins et al., 1996; Pendas et al., 1997).

THE MATRIX METALLOPROTEINASE FAMILY

9

Hemopexin Domain This domain of about 200 residues contains four repeats t h a t resemble hemopexin and vitronectin. There is a Cys residue at either end; these join and the resultant domain folds into a four-bladed propeller structure. All MMPs except MMP-7 contain this structure, but it does not appear to be essential for catalytic activity. Many truncated forms of MMPs have been produced t h a t lack this domain and all retain activity. However, substrate specificity for macromolecules may be greatly affected (see review by Murphy and Kn~iuper, 1997). The binding of TIMP is also assisted by this domain and, in the case of gelatinases, there is binding of TIMP to this domain even when the enzymes are in their zymogen form. Membrane Insertion Extension Although most MMPs have their C terminus at the end of the hemopexin domain, the four MT-MMPs have a further extension t h a t governs insertion of these proteases into the cell membrane. Its length ranges from about 80 to 110 residues. The membrane-spanning region of about 20 residues is about 20 residues in from the C terminus, leaving this short segment within the cytoplasm. MMP-19 has a 36-residue extension beyond the hemopexin domain but this is not membrane inserted. V.

THE BIOLOGICAL ROLE OF THE MATRIXINS

AO Do the M M P s Play a Major Role in Degrading the Extracellular Matrix? Table I provides a good overview of the vast number of biologic and pathologic processes in which it is believed the MMPs play an important, or even indispensable, role. However, it must be admitted that firm proof for such involvement is very sparse. The general sorts of proofs t h a t are offered include the use of in situ hybridization to show t h a t mRNA for a given MMP is present at the site of tissue remodeling, use of immunohistochemistry to demonstrate t h a t the enzyme protein is present, and localization of specific degradation products at the site (e.g., the cleaved fragments of collagen). These criteria, however, merely indicate guilt by association; they do not generally prove t h a t a specific MMP is responsible for a specific effect. Now t h a t we know of 17 MMPs in humans, it must be admitted that no one has examined a particular case of matrix remodeling for each activity nor established the contribution of each enzyme to the process. Although one can demonstrate the cleavage of collagen type II in arthritic cartilage (Dodge et al., 1991), for example, we now know that the identical specific cleavage can be produced by MMP-1, MMP-2, MMP-8, MMP-13, and MMP-14 and there is some evidence that each of these enzymes can be produced by chondrocytes.

10

J. FREDERICK WOESSNER, JR.

Krane (1994) has suggested criteria for proving the role of an MMP in a remodeling process: 9 Remodeling can be blocked with a drug or antibody specific to the MMP. 9 Remodeling can be reproduced by overexpression of the MMP gene in transgenic animals. 9 Remodeling can be abolished by deleting the MMP gene. 9 Spontaneous mutations can be identified and the phenotypes characterized. 9 Mutations can be induced in the gene t h a t reproduces the remodeling process. Progress in carrying out this program of establishing the roles of the various MMPs has been slow and somewhat disappointing. Shapiro (1997) has recently reviewed the results with transgenic m i c e ~ t h e r e are knock-outs for MMP-3, MMP-7, MMP-9 and MMP-12 and overexpressers of MMP-1 and MMP3. In most cases these have not led to sharply defined phenotypes and they have not led to death of the organism. A problem appears to be the high redundancy of function of the MMPs, so t h a t when one enzyme is knocked out another becomes more highly expressed to compensate for the loss. However, Shapiro has established quite clearly t h a t the mouse macrophage, with a somewhat limited repertoire of MMPs, requires MMP-12 (macrophage elastase) in order to produce emphysema in mice induced to inhale cigarette smoke ( Hau tama ki et al., 1997). A similar problem arises in the use of specific inhibitors to block the action of an MMP. In general, the active centers and bond specificity of the various MMPs are fairly similar. For example, the collagenase substrate D N P - P r o - L e u G l y * I l e - A l a - G l y - P r o - D - A r g is cleaved by all the MMPs tested to date. Furthermore, MMP inhibitors currently in use such as B a t i m a s t a t are hydroxamate compounds; these are found to inhibit members of the M10 family, such as TNF-a convertase, as well as MMPs (Black et al., 1997). However, progress is being made in making such inhibitors ever more specific as can be seen in Chapters 10 and 11.

B.

W h a t A r e the True S u b s t r a t e s for E a c h M M P ?

It is very common to find reviews containing a table similar to Table II in which there is a list of substrates for each MMP. A still more extensive compendium of substrates is provided by Chandler et al. (1997). However, I have avoided this practice because it is my opinion t h a t almost nothing is known about the na t ural substrates of the MMPs and t h a t such a table is very misleading. Investigators tend to test

THE MATRIX METALLOPROTEINASE FAMILY

11

only those matrix components ready to hand. Because many are cell biologists, their shelves contain collagens I and IV, fibronectin, laminin, and nidogen. However, I estimate t h a t there are more t h a n 100 known macromolecular components of the extracellular m a t r i x m a b o u t 30-40 each of collagens, proteoglycans, and glycoproteins. Examination of the digestion of this vast array of potential substrates by the 15 MMPs is barely under way. To establish a n a t u r a l e n z y m e - s u b s t r a t e relationship, it is necessary to show that the substrate is cleaved in vivo at a certain point that matches the specificity of the enzyme and is unlikely to be due to other proteases. The case on which the greatest attention has been lavished is that of the cleavage of the large aggregating proteoglycan aggrecan, found in articular cartilage. A good deal of circumstantial evidence was built up by showing that stromelysin 1 could readily digest the protein core of aggrecan and that stromelysin was elevated in osteoarthritic cartilage (Dean et al., 1989). However, more detailed study of the cleavage of aggrecan by various MMPs showed that every one of those tested, including collagenase 1, 2, and 3, could produce the same specific cleavage at the bond DIPEN*FFGVG (Fosang et al., 1996). Moreover, when aggrecan degradation products were isolated from cartilage and synovial fluid, the products had arisen from the cleavage of an ITEGE*ARGSV bond (Sandy et al., 1992). It therefore appears that a novel protease, aggrecanase, is required for this cleavage and that this is a metalloprotease but probably not an MMP. This case is sufficient to illustrate the problem. The redundancy of specificity of the MMPs makes it difficult to pin down which one did what, and very few cases have had the cleavage sites of the substrate examined in vivo. Even in a clear case such as the interstitial collagens, which show the specific cleavage by collagenases of their Gly-Ile bond, a cleavage unlikely to be produced by other proteases in the tissue, one is left with five contenders for the role of cleaving enzyme. However, in the transgenic mouse in which this collagen cleavage site is mutated, collagen is still degraded through cleavage of the telopeptides. In this case, MMP-13 appears to be the enzyme t h a t is capable of this cleavage (Krane et al., 1996).

C.

W h y So M a n y M M P s ?

In view of the redundancy of function of many of the MMPs, one may ask why so many enzymes have evolved. In spite of the difficulty of proving what each one does, I believe that the multiplicity of forms underlines the extreme importance of the MMPs for the normal morphogenesis, maintenance, and repair of the matrix. The individual

12

J. FREDERICK WOESSNER, JR.

cell is very highly dependent on its battery of MMPs to control its environment, to move through it, and to maintain its protective cocoon. The knock-outs seem to tell us that the cell can make do without one of these enzymes by mobilizing one or more of the remaining MMPs to take over the function of the deficient enzyme. Furthermore, we should not think that each cell possesses the entire battery of enzymes. So far as is known, enamelysin (MMP-20; Bartlett et al., 1996) is found only in developing tooth enamel. MMP-8 and MMP-12 are largely confined to the neutrophil and macrophage, respectively. MMP-7 appears to be restricted to epithelial cells (Wilson and Matrisian, 1996). The MMPs must be considered one more example of the profligacy of nature. VI.

WHAT THE FUTURE MIGHT HOLD

The imminent arrival of the millenium prompts one to prognosticate about the prospects for the MMP field. Activity in this area of research has now reached a fever pitch and, with the clear involvement of the MMPs in such major disease problems as cancer, arthritis, and atherosclerosis, it seems unlikely that a sudden decline in interest in the MMPs will occur. Diseases will remain a major driving force because economics and politics will favor the distribution of resources in that area. Current work on the development of specific inhibitors will intensify. Both the working out of the detailed specificity requirements of each MMP and the development of highly specific inhibitors will be facilitated by the increasing use of combinatorial chemistry. However, there appear to be limits to the specificity that might be achieved, with consequent unintended side effects, so increasing attention will be focused on regulating MMP activity through genetic techniques such as antisense methods and through the use of specific drugs/factors that can regulate the expression of each MMP. This will require more detailed knowledge of promoters and cell factors governing expression. Intervention may also be attempted at the level of proteolytic activation of the MMP zymogens. With respect to fundamental problems that remain to be answered, I am interested in the binding of MMPs to the cell and substratum. It is very difficult to extract most of the MMPs from tissues due to various types of anchoring. It is probably crucial for the cell to keep the MMPs in its vicinity to govern their activity, to keep track of how much enzyme is out there, and to prevent the enzymes from washing away with the blood until needed. A number of the MMPs appear to be attached to the cell surface bound to receptors, inserted into the membrane, or localized to invadopodia. This permits the cell to effect proteolysis in a specific direction following a regulated process of surface activation.

THE MATRIX METALLOPROTEINASE FAMILY

13

M u c h m o r e n e e d s to be l e a r n e d a b o u t h o w t h e cell s e n s e s t h e p r o t e o l y t i c a c t i v i t y in t h e o u t s i d e w o r l d - - t h r o u g h r e c e p t o r s , i n t e g r i n s , or s i m i l a r signaling mechanisms and through feedback from substrate fragments i n t e r a c t i n g w i t h t h e cell. W e n e e d to l e a r n m o r e a b o u t h o w t h e cell r e g u l a t e s t h e a c t i v a t i o n of M M P s a l r e a d y r e l e a s e d f r o m t h e cell. Finally, w i t h t h e d e t a i l e d s t r u c t u r e of M M P - T I M P c o m p l e x e s n o w in h a n d , we c a n b e g i n to e x p l o r e in m o r e d e t a i l h o w t h e T I M P s i n t e r a c t w i t h t h e e n z y m e s . T h e full s t o r y h e r e c a n only e m e r g e w h e n we see t h e full s t r u c t u r e of t h e g e l a t i n a s e s w i t h T I M P s in place on b o t h t h e p r o e n z y m e a n d t h e active e n z y m e . In s u m m a r y , t h e p a s t 36 y e a r s h a v e s e e n t r e m e n d o u s a d v a n c e s in o u r u n d e r s t a n d i n g of t h e s t r u c t u r e a n d a c t i v i t y of t h e M M P s . H o w e v e r , t h e a r e a s of i g n o r a n c e a p p e a r to be a l m o s t infinite in e x t e n t , p r o m i s i n g m a n y e x c i t i n g y e a r s a h e a d in t h e M M P field.

ACKNOWLEDGMENTS

The author is supported by grant AR-16940 from the National Institutes of Health. Dr. Neil D. Rawlings, The Babraham Institute, Cambridge, UK, generously provided Fig. 1.

REFERENCES

Bartlett, J. D., Simmer, J. P., Xue, J., Margolis, H. C., and Moreno, E. C. (1996). Molecular cloning and mRNA tissue distribution of a novel matrix metalloproteinase isolated from porcine enamel organ. Gene 183, 123-128. Barrett, A. J., Rawlings, N. D., and Woessner, J. F. (1998). "Handbook of Proteolytic Enzymes," Academic Press, London. Basset, P., Bellocq, J. P., Wolf, C., Stoll, I., Hutin, P., Limacher, J. M., Podhajcer, O. L., Chenard, M. P., Rio, M. C., and Chambon, P. (1990). A novel metalloproteinase gene specifically expressed in stromal cells of breast carcinomas. Nature 348, 699-704. Black, R. A., Rauch, C. T., Kozlosky, C. J., Peschon, J. J., Slack, J. L., Wolfson, M. F., Castner, B. J., Stocking, K. L., Reddy, P., Srinivasan, S., Nelson, N., Boiani, N., Schooley, K. A., Gerhart, M., Davis, R., Fitzner, J. N., Johnson, R. S., Paxton, R. J., March, C. J., and Cerretti, D. P. (1997). A metalloproteinase disintegrin that releases tumour-necrosis factor-alpha from cells. Nature 385, 729-733. Bode, W., Reinemer, P., Huber, R., Kleine, T., Schnierer, S., and Tschesche, H. (1994). The X-ray crystal structure of the catalytic domain of human neutrophil collagenase inhibited by a substrate analogue reveals the essentials for catalysis and specificity. E M B O J. 13, 1263-1269. Chandler, S., Miller, K. M., Clements, J. M., Lury, J., Corkill, D., Anthony, D. C., Adams, S. E., and Gearing, A. J. (1997). Matrix metalloproteinases, tumor necrosis factor and multiple sclerosis: An overview. J. Neuroimmunol. 72, 155-161. Cossins, J., Dudgeon, T. J., Catlin, G., Gearing, A. J., and Clements, J. M. (1996). Identification ofMMP-18, a putative novel human matrix metalloproteinase. Biochem. Biophys. Res. Commun. 228, 494-498.

14

J. FREDERICK WOESSNER, JR.

Dean, D. D., Martel-Pelletier, J., Pelletier, J.-P., Howell, D. S., and Woessner, J. F., Jr. (1989). Evidence for metalloproteinase and metalloproteinase inhibitor imbalance in human osteoarthritic cartilage. J. Clin. Invest. 84, 678-685. Dodge, G. R., Pidoux, I., and Poole, A. R. (1991). The degradation of type II collagen in rheumatoid arthritis: an immunoelectron microscopic study. Matrix 11, 330-338. Fosang, A. J., Last, K., Kn~iuper, V., Murphy, G., and Neame, P. J. (1996). Degradation of cartilage aggrecan by collagenase-3 (MMP-13). F E B S Lett. 380, 17-20. Gross, J., and Lapiere, C. M. (1962). Collagenolytic activity in amphibian tissues: A tissue culture assay. Proc. Natl. Acad. Sci. USA 48, 1014-1022. Hautamaki, R. D., Kobayashi, D. K., Senior, R. M., and Shapiro, S. D. (1997). Requirement for macrophage elastase for cigarette smoke-induced emphysema in mice. Science 277, 2002-2004. Krane, S. M. (1994). Clinical importance of metalloproteinases and their inhibitors. Ann. N.Y. Acad. Sci. 732, 1-10. Krane, S. M., Byrne, M. H., Lemaitre, V., Henriet, P., Jeffrey, J. J., Witter, J. P., Liu, X., Wu, H., Jaenisch, R., and Eeckhout, Y. (1996). Different collagenase gene products have different roles in degradation of type I collagen. J. Biol. Chem. 271, 28509-28515. Lepage, T., and Gache, C. (1990). Early expression of a collagenase-like hatching enzyme gene in the sea urchin embryo. E M B O J. 9, 3003-3012. McGeehan, G., Burkhart, W., Anderegg, R., Becherer, J. D., Gillikin, J. W., and Graham, J. S. (1992). Sequencing and characterization of the soybean leaf metalloproteinase. Structural and functional similarity to the matrix metalloproteinase family. Plant Physiol. 99, 1179-1183. Murphy, G., and Kn~iuper, V. (1997). Relating matrix metalloproteinase structure to function: Why the "hemopexin" domain? Matrix Biol. 15, 511-518. Pendas, A. M., Kn~iuper, V., Puente, X. S., Llano, E., Mattei, M. G., Apte, S., Murphy, G., and LSpez-Otin, C. (1997). Identification and characterization of a novel human matrix metalloproteinase with unique structural characteristics, chromosomal location, and tissue distribution. J. Biol. Chem. 272, 4281-4286. Puente, X. S., Pendas, A. M., Llano, E., Velasco, G., and LSpez-Otin, C. (1996). Molecular cloning of a novel membrane-type matrix metalloproteinase from a human breast carcinoma. Cancer Res. 56, 944-949. Rawlings, N. D., and Barrett, A. J. (1995). Evolutionary families of metallopeptidases. Methods Enzymol. 248, 183-228. Sandy, J. D., Flannery, C. R., Neame, P. J., and Lohmander, L. S. (1992). The structure of aggrecan fragments in human synovial fluid. Evidence for the involvement in osteoarthritis of a novel proteinase which cleaves the Glu 373-Ala 374 bond of the interglobular domain. J. Clin. Invest. 89, 1512-1516. Shapiro, S. D. (1997). Mighty mice: Transgenic technology "knocks out" questions of matrix metalloproteinase function. Matrix Biol. 15, 527-533. Wilson, C. L., and Matrisian, L. M. (1996). Matrilysin: an epithelial matrix metalloproteinase with potentially novel functions. Int. J. Biochem. Cell Biol. 28, 123-136. Woessner, J.F., Jr. (1994). The family of matrix metalloproteinases [review]. Ann. N.Y. Acad. Sci. 732, 11-21. Yang, M. Z., Murray, M. T., and Kurkinen, M. (1997). A novel matrix metalloproteinase gene (XMMP) encoding vitronectin-like motifs is transiently expressed in Xenopus laevis early embryo development. J. Biol. Chem. 272, 13527-13533.

Interstitial Collagenases John J. Jeffrey Department of Biochemistry and Molecular Biology, Albany Medical College, Albany, New York 12208

I. Introduction II. Matrix Metalloproteinase-l: The Original Interstitial Collagenase A. Chemistry of MMP-1 B. Catalytic Activity of MMP-1 C. Activation of proMMP-1 D. Collagenase-3:MMP-13 and Rodent Interstitial Collagenase E. H u m a n Neutrophil Collagenase: MMP-8 III. "Nontraditional" Interstitial Collagenases IV. Summary and Future Perspectives References

I.

INTRODUCTION

Since interstitial collagenases were last reviewed by this author in this series (Jeffrey, 1986), the area of interstitial collagenase chemistry and biology--in parallel with that of the matrix metalloproteinases in g e n e r a l - - h a s undergone explosive growth. The emergence of our knowledge of the collagenases as functional proteins, as well as the information provided by molecular biological studies of these molecules, has provided q u a n t u m leaps in our knowledge of the structure, chemistry, enzymology, and biology of these enzymes. So much information has emerged in the last 8 to 10 years that it is impossible in a chapter such as this to thoroughly explore every aspect of the subject in the detail that one would want. Thus, in this chapter, massive selectivity was exercised in presenting information in any kind of useful fashion. Accordingly, studies that speak definitively of the overall chemistry or biology of the interstitial collagenases have been given preference over more circumscribed investigations of phenomena in individual tissue physiology or disease state pathologies. As a result, a great deal of excellent and important work has been neglected, for which the author appologizes. Constraints, unpleasant as they are, are nevertheless necessary to attempt to present a coherent overview of the state of the field. In the introduction to the previous review of interstitial collagenases, the point was made that the specification of the precise threedimensional structure of the connective tissues places unusual, perhaps 15

Matrix Metalloproteinases

Copyright 9 1998 by Academic Press. All rights of reproduction in any form reserved.

16

JOHN J. JEFFREY

unique, requirements on biology. The distances over which this precision is required to extend in tissues is the analogue of light years in physics, and the necessity of the maintenance of directionality ("up," "down," "left," "left," "right," "how far?") add a further demand on biological information transfer. The intervening years have yielded little in the elucidation of these fundamental biological issues, but with the enormous increase in knowledge of the number and nature of the molecules involved in these processes, some level of information regarding these questions can be expected by the time this subject is revisited. This chapter intends to provide the general reader with a broad overview of our knowledge of the nature of interstitial collagen degradation and the enzymes that catalyze it.

II.

MATRIXMETALLOPROTEINASE-I: THE ORIGINAL INTERSTITIAL COLLAGENASE

The first vertebrate collagenase both purified to homogeneity as a protein and cloned as a cDNA was that from h u m a n fibroblasts (Stricklin et al., 1977; Goldberg et al., 1986). This enzyme, bearing the designation matrix metalloproteinase-1 (MMP-1), has served as the prototype for all other interstitial collagenases. In view of that historical fact, the characteristics of MMP-1 are reviewed in some detail and are then used as a template for comparison with other interstitial collagenases.

A.

Chemistry of MMP-1

The enzyme is secreted from the cell as a pair of zymogens of Mr approximately 52,500 (Stricklin et al., 1977). The two secreted proteins apparently differ from each other only in that one is glycosylated and the other either less so or not at all. The precise chemistry of the glycosyl moiety in the glycosylated form of the protein has not been defined, nor has its biological function. All that is known at this time is t h a t there is no significant effect of glycosylation on enzymatic activity (Stricklin et al., 1978). In all other respects examined to date, such as chromatographic behavior, sedimentation equilibrium, and substrate specificity, the two forms of the proenzyme behave identically. Given the fact, however, that the presence of differentially glycosylated forms of MMP-1 has been observed in a variety of species (e.g., Roswit et al., 1983), it seems fair to speculate that some function--unknown as y e t m i s served by the presence of the glycosyl residues. The interstitial collagenases so far identified belong to a class of proteinases known as the matrix metalloproteinases (MMPs) (Woes-

INTERSTITIAL COLLAGENASES

17

sner, 1991). A number of excellent reviews of this ever-expanding gene family have been published (Sang and Douglas, 1996; Birkedal-Hansen et al., 1993; Dioszegi et al., 1995; Nagase, 1994), to which the reader is referred for further specific information. Briefly, MMP-1, and all members of the MMP family, including interstitial collagenases, share a very similar domain structure, as illustrated in Fig. 1 (BirkedalHansen et al., 1993), composed of a single peptide, a proenzyme domain, a catalytic domain, a short (usually 16-17 amino acids in length) "hinge" region, and a C-terminal domain resembling hemopexin (Li et al., 1995; Faber et al., 1995), originally identified as a heme-binding protein (Jenne and Stanley, 1987). All the members of the MMP gene family except one, matrilysin, contain a similar domain and, interestingly, a similar hemopexin domain has also been identified in the cellmatrix adhesion protein vitronectin (Hunt et al., 1987; Jenne, 1991). All members of the MMP family contain Zn 2§ at the catalytic site (Vallee and Auld, 1992; Birkedal-Hansen et al., 1993) and, in addition, require Ca 2§ for stability and activity (Seltzer et al., 1976; McKerrow, 1987; Zhang et al., 1997). The intrinsic zinc at the catalytic site is chelated by the imidazole nitrogens of three histidine residues together with the sulfur of a free cysteine, all located in the proenzyme domain (SanchezLopez et al., 1993). The amino acid motif containing the critical histidines, HexGHxxGxxH, is not restricted to the interstitial collagenases but is present in highly conserved form in all MMPs (Sang and Douglas, 1996). In addition, the amino acid sequence surrounding a cysteine residue also believed to be required for latency (PRCGVPD) is highly

FIG. 1. Representation of a prototype matrix metalloproteinase. All the interstitial collagenases contain variations of the domains indicated in this representation. (Adapted from Birkedal et al., 1993.)

18

JOHN J. JEFFREY

conserved in all members of the MMP family (Sang and Douglas, 1996). A second molecule of zinc has also been shown to be present in the enzyme (Bode, 1994; Lovejoy et al., 1994); this second zinc, however, is believed not to be involved in the catalytic activity of the enzyme but rather to play a structural role, serving to stabilize the enzyme. Similarly, Ca 2§ has long been implicated in the activity of MMP-1, and recent studies indicate that, like the second zinc atom, calcium is involved in stabilizing the structure of the enzyme, in both the catalyticand the C-terminal hemopexin domains (Li et al., 1995; Gomis-Ruth et al., 1996). A number of X-ray crystallographic studies have now been performed on various forms of MMP-1 from a variety of species. The first complete crystallographic analysis of full-length MMP-1 was that of the porcine enzyme (Li et al., 1995), crystallized as a complex with a low-molecularweight inhibitor. For the purposes of this discussion, the structure presented consists of three important domains: (1) the amino-terminal catalytic domain, which, not surprisingly, bears strong resemblances to the catalytic domains of other MMPs crystallized to date; (2) the linker region, rich in proline, at residues 261-277, highly exposed to the environment; and (3) the hemopexin domain, revealed for the first time as four antiparallel beta-sheet subdomains, arranged in a four-bladed propeller structure stabilized by one of the disulfide bridges in MMP-1, that between Cys27s, which is at the very beginning of the domain, and Cys4~6, at the C-terminal end of the protein. A good feel for the geometry of the four-bladed structure can be obtained from the stereo Ca trace diagram taken from Li et al. (1995). A number of studies have shown that the hemopexin domain is crucial to the nature of the catalytic activity of MMP-1. Truncated MMP-1, which lacks the hemopexin domain, displays no activity against native collagen, although it retains enzymatic activity against low-molecular-weight peptide substrates (Murphy et al., 1992; Knauper et al., 1993). Similarly, substituting the hemopexin region of MMP-3 (stromelysin) for the native domain of MMP-1 results in essentially total loss of collagenolytic activity (Sanchez-Lopez et al., 1993). Thus, the hemopexin domain is a critical specifier of the enzymatic specificity of MMP-1, but the reason for this remarkably effective specification is not clear. Preliminary modeling studies by Li et al. (1995) and Bode (1994) to assess the possibility that the hemopexin domain directs the favorable binding of triple helical collagen to the enzyme and of the scissile bond to the active site suggest that no particularly favorable binding motif exists. To compound this conundrum further, other workers have pointed out, on the basis of crystallographic studies, that the active site is situated

INTERSTITIAL COLLAGENASES

19

in the enzyme such that it would be predicted to be relatively inaccessible to a peptide bond in triple helical collagen. Thus, the active site of interstitial collagenase, as currently viewed, appears to be an inhospitable one for accommodating the triple helical collagen molecule for the purpose of catalysis. Nevertheless, cleavage not only occurs, but appears to occur in an apparently en bloc fashion. That is, all three chains appear to be cleaved in a given substrate molecule before the products of any cleaved molecule are released (Welgus et al., 1981a). In addition, no evidence exists for a preferential cleavage of one of the a chains of a heterotrimeric collagen substrate vis ~ vis another. In unpublished experiments, this author has incubated native type I collagen [(a1)2(a2)] in solution with a molar excess of MMP-1 at low temperature to allow binding and then examined the ratio of cleavage products of the chains in the substrate at short times of incubation. In such a situation, it was hoped that if a preference for one chain versus another existed, the ratio of the three-quarter-length products derived from the a chains of the parent molecule would differ from 2:1, depending on what preference exists. No deviation from a 2:1 ratio was ever observed. Again, this result is consistent with the presence of a functional en bloc cleavage of collagen by MMP-1. It also appears, from the use of type I collagen mutated at the collagenase cleavage site in only the al chain, that the cleavage of the normal a2 chain in the same molecule cannot be catalyzed (Krane et al., 1996). Furthermore, a preliminary report (Byrne et al., 1992) indicates that the cleavable bonds in the a chains of a triple-helical collagen molecule must be in their native register for any cleavage to occur. That is, if the cleavable bond of one chain is moved with respect to the analogous bonds in the other two chains, no collagenase cleavage is observed. In other words, all three chains must be cleavable, or no chain is cleaved. It is, therefore, not easy to readily visualize a mechanism to accommodate all of these data, simply from the structural knowledge of MMP1 available at this time. An intriguing hypothesis has been presented by DeSouza et al. (1996), who have identified a sequence motif in the hinge region ofMMP-1 (GRSQNPVQPIGPQTP that theoretical considerations suggest would, by virtue of its repetitive PXX sequences, adopt a helical configuration similar to that of the collagen helix itself: a polyproline II-like helix. Based on this conformational analysis of the hinge region sequence, these authors propose that this region interacts with the collagen molecule at or near the cleavage site. This interaction is hypothesized to destabilize further the cleavage site area, long proposed to possess a "looser" helical structure as a reason for its collagenase susceptibilility (Miller et al., 1976). In the model of DeSouza et

20

JOHN J. JEFFREY

al., a possible induced relaxation of the substrate helix occurs, allowing only one chain of the triple helix to become available to the active site of the molecule, rather t h a n the difficult achievable fit of all three chains. In addition, this hypothesis contains a tacit imbedded principle: that of multidomain-induced conformational changes resulting in a transition to an ultimately favorable environment for collagenolysis. Although this proposal remains theoretical at present, it should provide the basis for further experiments to elucidate the processes whereby interstitial collagenases such as MMP-1 are able to cleave their substrate at such a specific locus and in such a unique way.

B.

Catalytic Activity of MMP-1

This general topic has been the subject of numerous studies and reviews, and it is not the purpose of this chapter to exhaustively revisit the details of these many investigations. Rather, it is intended to present some basic facts and features of MMP-1 activity and to attempt to put them in a rational biological context. MMP-1 degrades types I, II, III, and X collagens (Welgus et al., 1981a), type I gelatin, the antiprotease al-antitrypsin as well as its own zymogen in at least two places (see Birkedal, 1993, for further detail), and it is likely that other substrates exist for this enzyme as well. Its activity on the interstitial collagens has been most extensively studied; the enzyme cleaves the triple helix of types I, II, and III collagen exclusively at the Glym-Leu/Ileu776 peptide bond, very close to threequarters of the distance from the amino terminus of the substrate molecule (Miller et al., 1976). Aside from the considerations discussed in the previous section with respect to structural features of the enzyme molecule with potential relevance to catalytic activity, the nature of the cleavage site in native collagen that renders it susceptible to catalysis is essentially unknown. It has been proposed that the helix is weaker in this region by virtue of a low average hydroxyproline content compared to elsewhere in the collagen molecule, but no experimental evidence exists to confirm this hypothesis. If indeed the helix is less rigid in this area, then this region should be more susceptible to proteolysis. However, a wide array of enzymes of varying classes fails to cleave any bond in this area of the molecule, and MMP-1 cleaves only one of these bonds, with no evidence of"nibbling" in the hypothetically weak helix area (Welgus et al., 1980; Welgus et al., 1981a). One situation, that of the cleavage of type III collagen in some species (the h u m a n is one), is particularly instructive. Trypsin makes a single cleavage in these collagensmalthough it should be emphasized that

INTERSTITIAL COLLAGENASES

21

this cleavage is never observed in type I collagen~at an Arg-Gly bond, eight residues C-terminal to the collagenase cleavage site (Miller et al., 1976). Although this experimental finding is consistent with the weak helix hypothesis, type III collagens from some species are not similarly susceptible to trypsin (Welgus et al., 1985)~chick type III, for example. The same bond cleaved by trypsin in type III collagen exists in these type III collagens as well as in the type I collagens of the same species, yet is not cleaved by trypsin in these molecules (Welgus, Burgeson, et al., 1985). Thus, one is left with a murky picture of the requirement for substrate determinants of cleavage. One finding does stand out in this unclear area, however. Krane and colleagues (1996) have shown, as mentioned earlier, that all three chains of a collagen molecule must be cleavable to allow cleavage of any chain and that the three scissile bonds must be in "native" register for cleavage to take place (Byrne et al., 1992). If the scissile bond of one chain is moved in relation to the analogous bonds in the other chains in a collagen molecule, cleavage is again inhibited. Thus, although the mechanistic implications of these data are not known, it is clear that stringent spatial and structural requirements must be met for collagenolysis to proceed. The dynamics of the action of MMP-1 on various collagenous substrates was dealt with at length in the previous review of this subject by this reviewer (Jeffrey, 1986). To summarize, however, the following points can be useful: (1) MMP-1 cleaves all three classical interstitial collagens in solution, displaying classical Michaelis-Menten kinetics (Welgus et al., 1981a; Fields et al., 1987). Values for Km ranged from 1 to 2 tLM, indicating a rather high affinity of enzyme for substrate. The trade-off for this high affinity appears to be a very slow catalytic rate: the Vmaxfor the activity of MMP-1 on type I collagen in solution at 25~ is approximately 25 mo1-1 hr -~ (Welgus et al., 1981a). This value translates to the cleavage of a single collagen a chain per minute per enzyme molecule, a very low rate indeed in enzymology, especially considering the high affinity of the enzyme for its substrate. Of further interest was the large disparity in the rates at which MMP-1 degraded the three principal interstitial collagens. Type III was degraded some 10 times faster than type I collagen, and more than 100 times more rapidly than type II collagen (Welgus et al., 1981a). Indeed, type II was degraded so slowly that it gave rise to speculation that MMP-1 might not be involved in the degradation of cartilage collagen at all. This latter speculation turned out to be a fortunate guess, with the discovery of MMP-13 and its presence in cartilage together with MMP8 (vide i nfra). The interaction of MMP-1 with its various collagenous substrates has been the subject of extensive investigation, and the picture that

22

JOHN J. JEFFREY

emerges implicates the availability of water to the site of peptide bond hydrolysis as a critical determinant in the catalytic efficiency of the enzyme (Welgus et al., 1981b; Jeffrey et al., 1983). As previously described, the exclusion of water from collagenous substrates as they gain higher levels of organization massively affects the rate of collagen degradation by MMP-1. Thus, when this enzyme is examined for its ability to degrade gelatin chains, even though its catalytic activity (i.e., Vmax)is rather low, it behaves energetically as a "normal" enzyme (Welgus et al., 1981b). That is, it displays a temperature dependence characteristic of most enzymes in biology, with a so-called Q10 of approximately 2. In other words, for a change in reaction temperature of 10~ the rate of catalysis changes approximately 2-fold, indicating an activation energy (EA) of approximately 10 kcal/mol. When, however, the cleavage of the same chains as part of a native, triple helical collagen molecule is assessed, the energetics ofcollagenolysis change drastically. In this setting, the rate of collagenolysis varies much more drastically with temperature, and EA, as referred to earlier, rises to values in the vicinity of 40 kcal/mol. Such values specify a change in reaction rate of approximately 10-fold (as opposed to 2-fold for a normal enzyme) for every 10 ~ of temperature change. When the triple helical molecules are aggregated into native fibrils, the energetics change dramatically once again. The activation energy of MMP-1 on fibrillar type I collagen is of the order of 100 kcal/mol, which specifies a change of some 200-fold for every 10 ~ of temperature change, an unprecedented situation in enzymology. This value indicates that the rate of collagenolysis of fibrillar type I collagen triples for every 2 ~ of rise in temperature! Another remarkable phenomenon relating the availability of water and the collagenolytic activity of MMP-1 on fibrillar collagen is observed when D20 is substituted for H20 in the incubation mixture. Instead of the usual reduction in reaction rate of 2-fold or less, the rate decreases by 10-fold or more. Because this value cannot be explained in terms of classical isotope effects, it was hypothesized that this remarkable D20 effect was another example of the difficulty of getting water to the site of peptide bond hydrolysis (Jeffrey et al., 1983). Thus, the overall conclusion from a number of studies bearing on this situation was that the availability of the water required for the hydrolysis of the scissile peptide bond in collagen becomes more and more the rate-limiting step as the molecule becomes more organized. The aggregation of monomers to form the mature collagen fibril, with the exclusion of water that accompanies this process, presents the most difficult substrate of all for MMP-1. A diagrammatic representation of this process, as derived from the studies described, is illustrated in Fig. 2.

INTERSTITIAL COLLAGENASES

23

H-o,H

Peptide bond freely accessible to water

H

H

H H

. . . . .

H H

H H

H H

H H

o'

'o'

H H

MONOMER

~

FIBRIl.

m

H H "O" m m m

H

H

o

,o.

H

H H "O" m m ~

H

/ Peptide bond accessible to water with difficulty

H H "O" m

,'J"'"

I

/ Water almost completely excluded from interior of fibril

FIG. 2. Representation of the relationship of the availability of water to the ease of peptide bond hydrolysis by interstitial collagenase in collagenous substrates at progressive levels of organization. As the level of organization increases, the difficulty of bringing water to the site of hydrolysis increases drastically. [From Jeffrey, J.J. (1986). The biological regulation of collagenase activity. In "Regulation of Matrix Accumulation" (R.P. Mecham, ed.), pp. 53-92, Academic Press, New York, with permission.]

24

JOHN J. JEFFREY

When the data describing the energetics of MMP-1 activity are taken together with the considerations presented by the structural analyses of MMP-1, it is tempting to paint a picture of a combination of scissile bonds in a hydrophobic environment, requiring energy to transport the water of hydrolysis to the correct site. Together with major induced-fit changes in the enzyme molecule to fully enable an active site that appears to have difficulty accessing its substrate the result is a low rate ofproteolysis. Given the fact that processes of development, metamorphosis, morphogenesis, repair, and involution take place over long periods of biological t i m e - - a l l requiring spatial and temporal precision to ensure correct outcomes--it may be that nature has accepted these barriers to collagenolysis as an appropriate price to pay for spatial precision.

C. Activation of proMMP-1 Studies of MMP-1 as a protein and as an enzyme revealed that proMMP-1 can be activated by a wide array of enzymes, reagents, and manipulations. Thus, for example, enzymes as different as trypsin, kallikrein, cathepsin G, and plasminogen were shown by a number of laboratories to activate the zymogen to an active form of the enzyme, with an accompanying loss of about 10 kDa in mass (see BirkedalHansen et al., 1993, for a more comprehensive discussion). At the same time, a variety of apparently unrelated small molecules were found, as a group, to accomplish proMMP-1 activation as well. These included KSCN and KI at high concentrations (-3M), sodium dodecyl sulfate (SDS) and, most interestingly, a number of congeners of a series of organomercurial compounds, including phenyl mercuric chloride, phydroxy mercuribenzoate, and aminophenyl mercuric acetate. The earliest study to address the nature of the activation of the proenzyme by organomercurials strongly suggested that a conformational activation was occurring in the presence of these compounds (Stricklin et al., 1983). Thus, it was proposed that the initial active form of the enzyme was that of an "active zymogen," and that this conformationally active form subsequently underwent autolytic cleavage to the truncated active form. Because the autolytic process was independent of the concentration of enzyme protein, it was further proposed that this cleavage was an intramolecular, rather than an intermolecular process. In the intervening years, this hypothesis has been largely confirmed, in more specific terms, as the sequence of the molecule has become available. Cleavage sites of the various proteases that activate the zymogens have been identified. They are clustered at the center of the pro-piece (residues 35-40). In a sense, this region acts in a manner analogous to the "bait" region of a2-macroglobulin and allows a number of unrelated proteinases to catalyze this initial cleavage (see Birkedal-Hansen et

INTERSTITIAL COLLAGENASES

25

al., 1993, for more detail). Following the action of a proteinase, the resultant truncated zymogen is able to cleave the remainder of its own pro-piece, resulting in the production of the m a t u r e form of the enzyme. The ability of the truncated zymogen to catalyze an intramolecular cleavage has been attributed to the interruption of the C y s - Z n 2§ bond, giving rise to the oft-invoked "cysteine switch" mechanism of activation, wherein the switch is closed when the bond is intact and open when the bond is disrupted, allowing proteolysis to occur (Van Wart and Birkedal-Hansen, 1990). Attractive as this hypothesis is, it has not been unequivocally demonstrated to be operative, and some existing data cannot be fully explained by the universal participation of the cysteine switch mechanism in proMMP-1 activation. The most difficult data to reconcile with this mechanism come from reports from Chen et al. (1993) and Birkedal-Hansen's laboratory (Galazka et al.,1996), in which the cysteine in proMMP-3 (i.e., stromeylsin) putatively involved in the cysteine switch mechanism was m u t a t e d to a variety of other amino acids. Under these conditions, the p r o e n z y m e ~ a b s e n t the c y s t e i n e - - r e m a i n e d latent and, more importantly, could still be activated by organomercurials. These data are not consistent with the existence of cysteine chelated to the active site zinc as a primary regulator of proMMP-3 latency, and these investigators speculated that the crucial effect of the chemical activators of MMPs is the disruption of a salt bridge between other amino acids of the pro-piece and the active portion of the molecule (Galazka et al., 1996). Unfortunately, an identical study has not been done using proMMP-1, but it is of interest that, in the initial study of organomercurial activation of proMMP-1, Stricklin et al. (1983) were unable to detect alkylation of the sulfhydryl of the cysteine by p-chloromercuribenzoate (pCMB), a compound t h a t was effective in generating the "active zymogen" form of the enzyme. This compound has an extremely high affinity for free sulfhydryl groups (and for this reason was used for many years to determine the number of free cysteines in proteins); nevertheless, neither a well-established spectrophotometric method nor radiolabeling with [14C]-pCMB was successful in demonstrating alkylation of the - S H group of what we now know is the only unpaired cysteine in proMMP-1. All of these data, coupled with the ability of a number of chemically unrelated compounds to produce conformational activation, suggest the possibility t h a t this process is not as simple as once thought. What these compounds do have in common, as a group, is t h a t they all are chaotropes. That is, they all display the ability to disrupt the structure of water and, as such, might be expected to exert powerful effects on the conformation of the structure of proteins. Thus, it is possible that conformational distortion alone, without the need for a direct rupture

26

JOHN J. JEFFREY

of a co-ordinate covalent bond such as Cys-Zn 2§ is sufficient for the conformational activation of proMMP-1 and possibly other MMPs as well. This general question of proenzyme activation is raised in view of the fact that the precise mechanism of in vivo activation of proMMP-1 has never been established in any biological setting. A number of suggestions have been made, for example, that the plasminogen cascade serves as a major physiological activator of interstitial collagenase (He et al., 1989), but it is important to bear in mind that these proposals are based on test tube observations, and because the actual in vivo form of activated MMP-1 is unknown, the possibility of multiple pathways of activation must be entertained. This might even include conformational activation in biology: A report in the literature (Tyree et al., 1981) describes the apparent existence of proteins in conditioned culture medium of cells or tissue fragments which display the characteristics ofconformational activators of h u m a n proMMP-1. Thus, in the presence of these "factors," the proenzyme displays full activity at the time of addition of the activating factor, and examination of the resultant fully active enzyme shows that it retains the molecular weight of the zymogen. Unfortunately, purification of these putative proteins has not been further documented, so it is unclear whether they do indeed serve the same function in biology. Nevertheless, the possibility that nature makes use of the propensity of the MMPs to undergo conformational activation cannot be ruled out. To further illustrate the potential complexity of the processes of MMP-1 activation, the example of the potential participation of stromelysin (MMP-3) in the activation of proMMP-1 is instructive. Stromelysin is capable of catalyzing a further cleavage in the initial autocatalytically truncated form of the proMMP-1 as initiated, for example, by trypsin, resulting in a "mature" enzyme slightly different from that produced entirely by autolysis (Suzuki et al., 1990). The resulting product (FVL-collagenase) is nearly 10 times more active than the autolytically derived product: VL-collagenase (Suzuki et al., 1990). This remarkable change in the activity of the enzyme, as a result of the action of stromelysin, suggests the possibility that, in nature, a given amount of collagenase can be manipulated to exhibit a wide range of activity. As an example, in preliminary cell culture experiments (Jeffrey, unpublished observations) the production ofstromelysin by h u m a n fibroblasts clearly modifies the specific activity of MMP-1, again over a range of at least 10-fold. Of potentially more biological importance, however, is our observation in this system (Jeffrey, Ehlich, and Roswit, unpublished) that the inhibition of the production of stromelysin is some 10fold more sensitive to glucocorticoids than is the analogous inhibition of MMP-1 itself by these steroids. In such a setting, the activity of a

INTERSTITIAL COLLAGENASES

27

given amount of MMP-1 can be imagined to be regulated over a very wide range by such regulatory molecules as steroids, depending on the relative amounts of stromelysin and collagenase in a given biological setting. Thus, it is attractive to hypothesize that the collagenolytic potential in a given tissue can be regulated significantly by mechanisms that are designed to modify the specific activity of a constant amount of collagenase. The main point of the foregoing discussion is that the general question of how MMPs are activated in specific settings in vivo remains perhaps the major unknown area of matrix metalloproteinase biology today. The answer to this question presents a formidable challenge to the field, given the low levels of expression of an enzyme such as MMP-1 in general, and the likelihood of the focal nature of that expression, the determination of the characteristics of the true native active form of the enzyme will be difficult indeed. For example, the amount of collagen degraded in the uterus of a rat during postpartum involution is approximately 0.1 tLmol. Clearly the amount of collagenase, acting catalytically as it does, must be considerably less, presenting a massive challenge to our present methodology of protein characterization.

DQ Collagenase-3: M M P - 13 a n d R o d e n t I n t e r s t i t i a l Collagenase As indicated earlier in the section devoted to MMP-1 specificity, the latter enzyme was found to have such a low activity on type II collagen that its usefulness as an enzyme that could manage the collagen phenotype of cartilage was called into question (Welgus et al., 198 la). Indeed, it was hypothesized that a cartilage-specific collagenase might exist for this particular biological purpose. At the same time, an interstitial collagenase was purified from rat myometrial smooth muscle cells (Roswit et al., 1983). This enzyme catalyzed cleavage at the 3/4:1/4 site of the major interstitial collagens, but with significant quantitative differences from the cleavages catalyzed by human MMP-1 (Welgus et al., 1983). Although the rat enzyme displayed similar affinities (Km) for collagen substrates, the massive differences in Vmax displayed by human MMP-1 were not evidenced in the activity of the rat enzyme. Thus, the values for Vmaxdispalyed by rat interstitial collagenase were essentially identical for the three major interstitial collagens. Of special interest was the fact that type II collagen was degraded as efficiently as were types I and III. Recent observations (Jeffrey, Wilcox, Hambor and Mitchell, unpublished) suggest that this enzyme is apparently the only interstitial collagenase in the rodent genome. If this is indeed the case, the equal efficiency of cleavage by the enzyme of all the interstitial

28

JOHN J. JEFFREY

collagens would allow this MMP to effectively manage collagenolysis in a wide variety of tissues. Interestingly, the rat enzyme displayed high levels of gelatinolytic activity, in marked contrast to h u m a n MMP-1. Although the two enzymes displayed similarities in the bonds cleaved in denatured collagen substrates (Gly-Leu, Gly-Ileu, Gly-Phe, Gly-Ala), the rat enzyme displayed a distinctly higher specific activity against gelatin than against native collagen, while the reverse was true for human MMP-1 (Welgus et al., 1985). In 1990, the rat enzyme was cloned (Quinn et al., 1990), and it became clear that this collagenase and h u m a n MMP-1 were very different in homology (less than 50% similarity exists), although the major modules characteristic of MMPs (Quinn et al., 1990; Sang and Douglas, 1996) were maintained. Furthermore, this enzyme appeared to be produced by a variety of mesenchymal cells in the rat--osteoblasts, fibroblasts, and smooth muscle cellsmindicating that the enzyme was not specialized to a single cell type (Quinn et al., 1990). This difference between rodent and h u m a n collagenases remained a puzzle, especially when mouse interstitial collagenase was cloned and found to be essentially identical to the rat enzyme (Henriet et al., 1992). By contrast, the interstitial collagenases cloned from rabbit and pig sources displayed much higher homology to h u m a n MMP-1 than to the rodent collagenases (see Sang and Douglas, 1996, for detailed homology comparisons). Finally, Freije et al. (1994) cloned a h u m a n homologue of rodent collagenase from a breast carcinoma cDNA library. The homology between this second h u m a n enzyme of resident cell origin and the rodent mesenchymal enzyme is nearly 90%, although some significant sequence differences nevertheless exist, particularly in the proenzyme domain. This enzyme has been designated as MMP-13 or, alternatively, h u m a n collagenase-3. Tissue distribution of MMP-13 was originally viewed as quite limited; indeed, initially it appeared to be specific to breast carcinoma cells in situ, but more recently it has become apparent that chondrocytes are a major source of this collagenase. Two groups have now identified MMP-13 in both normal and osteoarthritic articular chondrocytes (Mitchell et al., 1996; Reboul et al., 1996). Furthermore, the enzyme has been examined for collagen substrate specificity (Mitchell et al., 1996; Billinghurst et al., 1997; Knauper et al., 1996) and found to cleave type II collagen--the major interstitial collagen of cartil a g e - a t a much higher rate than types I and III. Thus, this enzyme appears to fulfill the role, only a matter of speculation 15 years ago, of a type II selective collagenase. This finding has understandably engendered considerable excitement among investigators attempting to understand the pathophysiology of rheumatoid arthritis and osteoar-

INTERSTITIAL COLLAGENASES

29

thritis, pathologic conditions in which the degradation of type II collagen has long been postulated to play a major role. Human MMP-13 is very similar, but by no means identical, to its rodent homologues (Sang and Douglas, 1996). By way of similarity, it appears that human MMP-13 displays the ability to cleave type II collagen at least as efficiently as types I and III. In the case of the rat homologue, the catalytic efficiency against type II collagen is approximately equal to that exhibited against types I and III. In the case of human MMP-13, the catalytic efficiency of the enzyme against type II collagen is some 10-fold higher than the efficiency of the enzyme on types I and II. Suprisingly, given the apparent universality of bond cleavage shared by interstitial collagenases, MMP-13 catalyzes an additional cleavage in the triple helix of type II collagen, one triplet aminoterminal to the primary cleavage site (Billinghurst et al., 1997). The biological significance of this unanticipated extra cleavage in cartilage collagen is unknown. In addition, it has been found that human MMP13 degrades denatured collagen extremely efficiently, as had previously been shown for rat collagenase. Indeed, the gelatinolytic activity of both rat collagenase and human MMP-13 is substantially greater than the coUagenolytic activity of either enzyme (Welgus et al., 1985; Mitchell et al., 1996). The full biological implication of this capability of MMP13 homologues is unclear at this time, but a report from Fosang et al. (1996) indicates that MMP-13 has the ability to cleave aggrecan, a possible indicator of the existence of yet additional substrates for this subfamily of enzymes. In further support of this notion, Krane and coworkers (1996) have shown that rodent collagenases, but not human MMP-1, have the ability to cleave the telopeptide region of type I collagen in addition to the classical 775-776 cleavage site. This activity appears to be determined by the amino-terminal domain of the rodent collagenases. Chimeric collagenases with MMP-1 N-terminal domains fail to catalyze this cleavage, whereas chimeras containing rodent (i.e., MMP-13) N-terminal domains do catalyze this extra cleavage in the collagen molecule. These authors suggest that this additional proteolytic capability of MMP-13 might allow some level of collagen degradation to proceed during development in the transgenic mouse. Thus, it is possible that enzymes of this subgroup of interstitial collagenases are designed to degrade a wider variety of substrates than MMP-1 homologues. In general then, the catalytic activities of both rodent collagenase and MMP-13 display the similarities one might predict from the extensive amino acid homologies they share. X-ray crystallographic studies of the C-terminal (hemopexin) domain ofMMP-13 complexed with an active site inhibitor have been performed (Gomis-Ruth et al., 1996). The structure of this domain in MMP-13 is

30

JOHN J. JEFFREY

very similar to that present in porcine MMP-1, taking the form of a fourbladed propeller consisting of four antiparallel beta-sheet domains. No difference could be identified as providing a likely basis for the difference in collagen substrate specificities of the two MMPs, even though this domain is critical for the activity of both enzymes. Thus, the structural determinants of the specificity of the interstitial collagenases remain elusive. The activation of MMP-13 homologues display marked similarities to the analogous processes involved in MMP-1 activation. Although a thorough examination of the ability of chaotropes to activate MMP-13 has not been performed, it is clear that aminophenyl mercuric acetate (APMA) is an effective activator of the human enzyme, whereas trypsin appears to degrade the enzyme (Knauper et al., 1996). In the rodent, on the other hand, our laboratory has consistently observed that trypsin is an extremely reliable activator of rat interstitial collagenase, while APMA appears to promote loss of activity by degradation of the enzyme protein (Jeffrey and Roswit, unpublished observations). A thorough study, under carefully controlled conditions, is clearly required to obtain an appropriate comparison of the two molecules. For now it is sufficient to conclude that they share far more similarities than differences. A noteworthy finding regarding human MMP-13 is that it is apparently activated by the membrane-bound MT1-MMP, a property not shared by human MMP-1 (Knauper et al., 1996). This finding suggests the possibility of localized, pericellular activity of MMP-13 by virtue of its ability to be activated by a protease associated with the cell membrane. This could be important in the setting of breast carcinoma, in which human MMP-13 was first observed. Somewhat strangely, new knowledge of the role of this protease in the tumorigenic processmwhether in breast or other carcinomasmhas not been forthcoming. The rodent MMP-13 homologues have been utilized to develop a fascinating and informative animal molecule, that bids fair to illustrate the in vivo role of interstitial collagenase in development as well as in normal adult life. Krane and colleagues have used a targeted mutagenic strategy to develop a mouse whose homozygotic offspring contain only type I collagen, which is mutated at the 775-776 cleavage site (Liu et al., 1995) of the a chains. Careful studies have revealed that his mutation completely prevents degradation of type I collagen by interstitial collagenases, of both human and rodent sources. These homozygotes readily conceive and deliver normal litters of apparently normal offspring. This phenomenon alone was surprising to many in the community, given the tacit assumption over the years that collagenase would be required for many of the events that constitute normal connective tissue development; such is apparently not the case.

INTERSTITIAL COLLAGENASES

31

On the other hand, phenotypic aberrations do occur in the homozygous mice: principal among them is the inability to properly involute the uterus postpartum. This massive, and relatively rapid, physiological removal of collagen has stood for years as a principal paradigm for the action of collagenolytic enzymes (Jeffrey, 1991). In the mutant mice, however, massive nodules ofunresorbed collagen remain in the myometrium for long periods after parturition, adding further evidence for the requirement for interstitial collagenase in post-partum uterine involution. In addition to this impressive consequence of the existence of nondegradable collagen, the homozygous mice slowly develop other connective tissue aberrations as they age. Abnormal thickening of the papillary dermis occurs, as does kyphosis of the spinal column. Last, contractures develop in limbs, which appear to be the result of the failure of tendons to be able, by virtue of normal remodeling processes, to accommodate the growth of bones during postnatal development (Krane, personal communication). These slowly developing connective tissue defects are hypothesized to be the result ofvery slow and/or very focal processes of interstitial collagenase activation and function. Such processes would be difficult to detect with current techniques designed to assess major processes of tissue collagen degradation, but are likely to represent the biologically significant role of interstitial collagenases in mammalian growth and development. It will be exciting to follow further phenotypic consequences of this crucial targeted mutation in mammalian type I collagen as they are explored in a variety of other tissues. The discovery of a human homologue of previously defined rodent collagenases has helped to both clarify and confuse our understanding of the strategy adopted by biology for the degradation of interstitial collagens. On the one hand, the finding of this enzyme in cartilage-both normal and pathologicmprovides the possibility of an answer to the long-awaited question of how the degradation and remodeling of type II collagen are managed in mammalian systems. On the other hand, the question of why species such as h u m a n - - a n d many others-have two mesenchymal interstitial collagenases, whereas rodents appear to manage essentially the same biological processes with only one, remains a paradox. Clearly, this area of interstitial collagenase biology will be a major focus of research in the future. Of particular interest will be further information on the involvement of MMP-13 in tumors, given that the original localization of this enzyme indicated that breast carcinoma cells were a major, and originally apparently the only, source of the enzyme in a first approximation of biological localization.

E. Human Neutrophil Collagenase: MMP-8 This enzyme represents the third major mammalian interstitial collag e n a s e ~ a l t h o u g h historically the second to have been described~and

32

JOHN J. JEFFREY

remains one of the most difficult for which to assign specific physiologic or pathologic roles. Neutrophil collagenase has the distinction of being the only interstitial collagenase to be stored in cellsmin this case in the specific granules of the neutrophilmrather than being synthesized and released on biological demand. This enzyme was first described in the late 1960s (Lazarus et al., 1968), purified in limited amounts in the 1980s (Hasty et al., 1987), and cloned and subsequently expressed as a pure protein in the early 1990s (Hasty et al., 1990). The results of all these studies clearly put neutrophil collagenase in the MMP family and establish it as a member of the interstitial collagenase family, with the designation MMP-8. Curiously, another circulating cell, the monocyte/macrophage, fails to produce MMP-8; rather it produces MMP-1 in an "on-demand" fashion and, unlike the neutrophil, does not store the enzyme (Wahl 1977; Welgus, Campbell, et al., 1985). As a protein, the neutrophil interstitial collagenase is almost identical in size to other interstitial enzymes as a protein, but is considerably more highly glycosylated. Thus, fully glycosylated, the proenzyme form of MMP-8 is approximately 60 kDa; activation results in the loss of approximately 20,000 Da of mass, and the active form of the enzyme is approximately 40 kDa. Again, in the case of the neutrophil collagenase, the function of the glycosylation is unknown. However, again in common with MMP-1, it appears not to affect catalytic activity (see Birkedal-Hansen, 1983, for further details). Neutrophil collagenase displays a substrate specificity profile different from that of MMP-1 (Hasty et al., 1987). The preference for type III relative to type I collagen, which exists for MMP-1, is reversed in the case of MMP-8, which digests soluble type I collagen at a significantly higher rate than type III. Of particular interest in this area is the ability of MMP-8 to cleave type II collagenmthe predominant collagen of cartilagemat a markedly higher rate than does MMP-1. The molecular determinants of these different specificities for two very similar interstitial collagenases remain to be elucidated. Of potential major significance, however, relating to the ability of MMP-8 to degrade type II collagen at appreciable rates is the recent finding that MMP-8 is produced by normal h u m a n articular chondrocytes in the presence of low concentrations of interleukin-1 (Cole et al., 1996). It is clear from this study that it was the chondrocyte, and not contaminating leukocytes, that produced the MMP-8 and that the chondrocyte produces the enzyme in an on-demand fashion rather t h a n storing it in granules as is the case in the neutrophil. Thus, clearly, the neutrophil is not the sole source for this interstitial collagenase, and it will be exciting to follow the emerging biology of this MMP in

INTERSTITIAL COLLAGENASES

33

the cartilage. Together with the presence of MMP-13 in cartilage (see earlier discussion), the potential for type II collagen degradation is considerable, particularly in pathologic states. In addition, although some controversy exists as to exclusivity, MMP-8 has been shown to possess the ability to degrade aggrecan, a major s t r u c t u r a l glycosaminoglycan in cartilage (Fosang et al., 1996; Arner et al., 1997). The cleavage specificity of MMP-8 on the interglobular domain of aggrecan appears to differ from the specificity of the major, although as yet unidentified, cartilage aggrecanase (Arner et al., 1997); nevertheless, the presence of this enzyme in osteoarthritis or r h e u m a t o i d arthritis m a y signify an even more i m p o r t a n t role for MMP-8 in the pathophysiology of these diseases. Table 1 is a selected compilation of kinetic data, comparing the ability of the three interstitial collagenases to degrade a variety of collagen types. Although not complete, the values in the table illustrate the notions discussed previously. By way of s u m m a r y , the catalytic efficiencies (kcat/Km) for h u m a n MMP-13 and h u m a n MMP-8 on type II collagen are massively higher t h a n t h a t of h u m a n MMP-1 on this cartilage collagen. This, together with the recent localization of both enzymes in chondrocytes d f u r t h e r suggests t h a t they have been adapted for h a n d l i n g the normal development and repair of cartilage

TABLE I SELECTED KINETIC PARAMETERS OF INTERSTITIAL COLLAGENASES ON COLLAGEN SUBSTRATES

Enzyme Human MMP-1 Human MMP-13 Rat MMP-13 Human MMP-8

Collagen substrate

Km(t~M)

gcat (h -1)

kcat/Km (p~M-lh -1)

Human I Human II Human III Human I Human II Human III Human I Human II Human III Human I Human II Human III

0.8 2.1 1.7 n.d. 2 n.d. 0.9 0.9 1.7 1 2 2.2

53 1.0 350 n.d. 23 n.d. 10.7 14.2 20.2 450 150 200

67 0.48 206 50* 11.5 n.d. 12.3 15.7 11.9 450 75 80

All values obtained from studies on collagens in solution at 25~ using the method of Welgus et al., 1981. *Calculated from the specific activity presented in Knauper et al., 1997, assuming a Km of ~1/zM.

34

J O H N J. J E F F R E Y

collagen; their presence also allows for significant untoward collagenolysis in pathological states as discussed previously. MMP-8 contains all the modules that have become hallmarks of matrix metalloproteinases: a proenzyme domain, followed by a catalytic domain, the hinge region, and then by the characteristic C-terminal hemopexin-like domain. The catalytic site and cysteine switch regions display high homology to the analogous regions of all the MMPs, and structure-activity relationships are similar to those displayed by the other interstitial collagenases. Thus, for example, C-terminal truncation results in the loss of triple helical degrading activity, but retention of activity on other substrates (again, see Sang and Douglas, 1996, and Birkedal-Hansen, 1993, for further detail). Of particular interest with respect to domain requirements for substrate specificity in the MMPs as a class, the neutrophil collagenase has been used rather extensively by two groups to examine this issue, particularly with respect to the involvement of the hinge region in determining collagenolytic activity. By using a variety of chimeric molecules containing deletions and/or swapped domains from stromelysin, it has been shown (Hirose et al., 1993) that a 62-amino-acid sequence in the hinge region of MMP-8 was required for collagenolytic activity. Substituting the analogous, although longer, sequence from stromelysin failed to restore activity. Knauper et al. (1997) have reexamined this same region. Using alanine-scanning mutagenesis techniques, these workers found that mutating four proline residues in the hinge region resulted in profound reductions in collagenase activity. The results of this study are in substantial concurrence with the earlier proposal by DeSouza et al. (1996), which is discussed previously in some detail. Briefly, this proposal hypothesizes that the hinge region of interstitial collagenase adopts a polyproline II-like structure which interacts in a critically important way with the constituent chains of native collagen substrates. It is possible that a destabilization of the substrate occurs, allowing collagenolysis to proceed in the protected environment produced by the trapping of the substrate between the hemopexin domain and the active site region of the enzyme. Undoubtedly, this exciting possibility will receive considerably more attention in future studies of structural requirements for collagenolysis. The activation of neutrophil collagenase, released as it is as a proenzyme upon neutrophil degranulation, has been the subject of considerable study over the years. Activation can be achieved in vitro by a number of proteases and organomercurials and by autolysis, at sites identical to or very similar to those observed for the same processes in MMP-1. In addition, a number of studies have strongly implicated the role of products of the myeloperoxidase pathway in activated neutro-

INTERSTITIAL COLLAGENASES

35

phils, such as hypochlorous acid and monochloroamines in the in vivo activation of the enzyme (Test and Weiss 1986; Saari et al., 1992). A number of in vitro studies support the ability of HOC1 and chloramines such as taurine to activate neutrophil collagenase. In addition, cathepsin G, released upon neutrophil activation, has been shown to activate the proenzyme (Capodici et al., 1989), and considerable discussion has ensued as to which pathway is significant in biology. Claesson et al. (1996) have provided evidence that both pathways may play a role. When neutrophil activation is performed under aerobic conditions, significant HOCl-mediated activation occurs quite rapidly. Subsequent to this initial activation a further activation occurs, most likely mediated by cathepsin G. Prevention of cathepsin G activity under these conditions has little effect on the initial enzyme activation. Under anaerobic conditions, however, although the magnitude of activation is reduced from aerobic levels, all the activation can be prevented by serine proteinase inhibitors. Thus, these authors suggest that the two mechanisms--HOC1/chloramines and proteolytic/cathepsin G--exist to provide for effective activation in both highly oxygenated and relatively anaerobic tissue milieu. The potential availability of multiple pathways of activation may be mirrored by the analogous setting observed with proMMP-1, in which so many pathways have been shown to activate the proenzyme effectively. It may be that here, as well as in the neutrophil, biology provides a form of parallel processing to ensure that sufficient enzyme will be made available under potentially variable tissue conditions. III.

"NONTRADITIONAL"INTERSTITIAL COLLAGENASES

The original central dogma of this area of extracellular matrix chemistry was that only interstitial collagenase could cleave the native collagen triple helix. A small, but significant tear in that seamless theory was provided by Miller and colleagues (1976), who found that trypsin and elastase could catalyze a single cleavage in type III collagen. High concentrations of these enzymes were required, and the finding was viewed more as a tool to characterize the region of cleavage in the substrate than as a harbinger of new enzymes that might cleave native collagen. Recently, however, two reports have appeared that convincingly indicate that two enzymes, MMP-2 and MT1-MMP, can catalyze the characteristic Gm-I/L77~ cleavage in type I collagen. Aimes and Quigley (1995) have reported that human MMP-2, or 72-kDa gelatinase, free of the TIMP-2 that is normally associated with this enzyme during its biosynthesis, catalyzes this characteristic cleavage in soluble calf and chick type I collagens and appears to degrade native fibrillar

36

JOHN J. JEFFREY

type I collagen as well. Interestingly, the homologous 92-kDa gelatinase (MMP-9), similarly TIMP-free, fails to catalyze these cleavages. The enzyme displays values of kcat that are quite close to those exhibited by MMP-1, the original interstitial collagenase. Values for Km for MMP2, with soluble collagen as substrate, are some five- to eight-fold higher; that is, the affinity of this enzyme is considerably lower than that displayed by other, so-called "classical" interstitial collagenases ( - 1 t~M). The significance of this disparity in biology is not clear. In earlier studies by Welgus et al. (1980) it was calculated that the effective concentration of collagen in fibrils, that is, that available to an enzyme in solution, is approximately 1 t~M. This derives from the observation that the enzyme has access only to the molecules on the surface of the fibril. Assuming a Kd of the enzyme for its substrate of approximately micromolar, only half of what one imagines to be quite a high concentration of enzyme in biological terms ( - 5 0 t~g/mL)would be bound to collagen fibrils under equilibrium conditions. Thus, an enzyme with a Kd eight-fold higher would be bound to a correspondingly lesser degree. The challenge for the future will be to determine whether there are biological settings in which the concentration of MMP-2 is indeed high enough to allow for significant binding to fibrillar substrates. In addition, the study of Aimes and Quigley (1995) indicates that only MMP2 free of its normally associated TIMP-2 is able to degrade native collagen. This finding is intriguing in that it suggests the possibility that directed biological mechanisms exist whereby the extent of the MMP-2/TIMP complex could be modulated in vivo. Again, future studies will be required to explore this possibility. Finally, it has been recently shown that two members of the membrane-bound MMP familymMT1-MMP and a truncated version of the enzyme lacking the transmembrane domain, produced in recombinant formmboth cleave native collagen, again into the classical 3/4:1/4 fragments (Ohuchi et al., 1997). Heretofore, these molecules have been thought to interact only with proMMP-2 and proMMP-13 for the purpose of activating these enzymes. The findings of Ohuchi et al. now provide the possibility that pericellular collagenolysis by these molecules could play a role in biology. Note that, in the case of full-length MT1-MMP, the values of Kin and Vmaxfor this enzyme on type I collagen are very similar to those displayed by MMP-1, indicating a considerable potential for collagenolysis at or near the cell surface. One experiment that was not performed in these studies was to examine the potential of native MT1-MMP, bound to membranes, to exhibit the same collagenolytic capability; only purified recombinant forms of the enzymes were studied. If the native membrane-bound form does, in fact,

INTERSTITIAL COLLAGENASES

37

exhibit the ability to degrade native collagen, it would provide an even more exciting setting for this family of proteases. IV.

SUMMARYAND FUTURE PERSPECTIVES

The growth in our knowledge of the chemistry and biology of the matrix metalloproteinases in general has been explosive in the last several years, and increases in our knowledge of the interstitial collagenases have paralleled t h a t growth. The identification, in just the last 3 or 4 years, of new interstitial collagenases and old enzymes with newly identified interstitial collagenolytic activities expands the biological repertoire of interstitial collagenolysis, requiring considerable reevaluation and experimental approaches to assessing the roles of these proteases in appropriate systems. Molecular biological technology has allowed for the cloning and expression of all the known interstitial e n z y m e s m b o t h "classical" and "nonclassical"mand has provided pure proteins as well as specific antibodies to the enzymes. This technology has allowed the identification of domains responsible for specific biochemical characteristics of the enzymes and of domains t h a t are crucial for one or another aspect of their function. Thus, the era of the "impure reagent" is past; with the tools now at hand investigators will be able to examine complex biological and pathological settings at a new, more precise level of approximation. X-ray crystallography has, in the past 5 years, become an essential tool in our understanding of not only the overall structure of the interstitial collagenases, but perhaps more importantly, the relationships between domains in the enzymes. Extremely valuable insights have been gained into the role of the C-terminal hemopexin domain, for example. It is fair to say that extension of these studies will shed yet more light on the specific mechanism by which the interstitial collagenases manage the cleavage of the chains of native triple helical collagen. Crystallographic studies have also been invaluable in facilitating the design of inhibitors that are selective for one MMP over another. The ultimate goal is to achieve specificity of inhibition of individual matrix metalloproteinases, both to elucidate better the role of each in a variety of physiologic processes and in disease states as well. Even at the state of the art today, there is promise for therapeutic value from some of the inhibitors available now; this situation is bound to improve in the next few years, an exciting prospect indeed. One crucial area of MMP biology yet to be e l u c i d a t e d m a n d the interstitial collagenases are no exceptionmis the definition of the mechanism or mechanisms by which the pro-form of the enzyme is activated. As noted earlier, this will be a formidable task, but one that neverthe-

38

JOHN J. JEFFREY

less would open n e w horizons, both in our u n d e r s t a n d i n g of the basic biology of the e n z y m e s a n d in our ability to m o d u l a t e t h e i r activity in r a t i o n a l ways. Finally, t r a n s g e n i c a n d t a r g e t e d m u t a t i o n a l technology, as typified by the studies of Liu e t al. (1995), should provide m a j o r i n s i g h t s into the roles of these e n z y m e s in d e v e l o p m e n t a l a n d pathologic processes. This especially p r o m i s i n g a r e a bids fair to provide m a s s i v e increases in our knowledge of the sequelae of slow, low-level, long-term processes of collagen d e g r a d a t i o n i n v i v o , w h e r e m u c h of the t u r n o v e r of collagen a p p e a r s to t a k e place. In closing, the following observation w a s m a d e concerning the a t t a i n m e n t of the u l t i m a t e goals in our u n d e r s t a n d i n g of collagen degradation: as unattainable as they may appear at this time, one must have a certain faith that, as the molecules involved in these processes continue to be identified, purified and described in precise chemical terms, we are ipso facto approaching these ultimate answers. Twenty-five years ago workers in this field despaired of even finding a true collagenolytic enzyme; the breadth and depth of developments since that time should encourage us to think that today's insoluble problems will bring a set of solutions as exciting as those that have appeared since that time. At the end of this chapter, as the d e v e l o p m e n t s in this field t h a t h a v e occurred since the mid-1980s are discussed, it is a p p r o p r i a t e to echo a n d r e e m p h a s i z e those s a m e t h o u g h t s a n d expectations. REFERENCES

Aimes, R.T., Quigley, J.P. (1995). Matrix metalloproteinase-2 is an interstitial collagenase--Inhibitor-free enzyme catalyzes the cleavage of collagen fibrils and soluble native type-1 collagen generating the specific 3/4 length and 1/4 length fragments. J. Biol. Chem. 270, 5872-5876. Arner, E.C., Decicco, C.P., Cherney, R., et al. (1997). Cleavage of native cartilage aggrecan by neutrophil collagenase (MMP-8) is distinct from endogenous cleavage by aggrecanase. J. Biol. Chem. 272, 9294-9299. Billinghurst, R.C., Dahlberg, L., Ionescu, M., et al. (1997). Enhanced cleavage of type II collagen by collagenases in osteoarthritic articular cartilage. J. Clin. Invest. 99, 15341545. Birkedal-Hansen, H., Moore, W.G.I., Bodeen, M.K., et al. (1993). Matrix metalloproteinases: A review. Crit. Rev. Oral Biol. Med. 4,197-250. Bode, W. (1994). The X-ray crystal structure of the catalytic domain of human neutrophil collagenase inhibited by a substrate analogue reveals the essentials for catalysis and specificity. E M B O J. 13, 1263-1269. Byrne, M.H., Wu, H., Birkhead, J.R., et al. (1992). Sliding the collagenase cleavage site in the alpha-l(I) chain out of phase with the alpha-2(I) chain confers collagenase resistance. J. Bone Min. Res. 7, s131.

INTERSTITIAL COLLAGENASES

39

Capodici, C., Muthukumaran, G., Amoruso, M.A. (1989). Activation of neutrophil collagenase by cathepsin G. Inflammation 13, 245-258. Chen, L.-C., Noelken, M.E., Nagase, H. (1993). Disruption of the cysteine-75 and zinc ion coordination is not sufficient to activate the precursor of matrix metalloproteinase3 (stromelysin-1). Biochemistry 32, 10289-10295. Claesson, R., Karlsson, M., Zhang, Y.Y., et al. (1996). Relative role of chloramines, hypochlorous acid, and proteases in the activation of human polymorphonuclear leukocyte collagenase. J. Leuk. Biol. 60, 598-602. Cole, A.A., Chubinskaya, S., Schumacher, B . , e t al. (1996). Chondrocyte matrix metalloproteinase-8: Human articular chondrocytes express neutrophil collagenase. J. Biol. Chem. 271, 11023-11026. DeSouza, S.J., Pereira, H.M., Jacchier, S., et al. (1996). Collagen/collagenase interactionmDoes the enzyme mimic the conformation of its own substrate? F A S E B J. 10, 927-930. Dioszegi, M., Cannon, P., and Van Wart, H.E. (1995). Vertebrate collagenases. Methods Enzymol. 248, 413-431. Faber, H., Groom, C.R., Baker, H.M., et al. (1995). 1.8 A crystal structure of the Cterminal domain of rabbit hemopexin. Structure 3, 551-559. Fields, G.B., Birkedal-Hansen, H., and Van Wart, H.E. (1987). Sequence specificity of human skin fibroblast collagenase. J. Biol. Chem. 262, 6221-6226. Fosang, A.J., Last, K., Knauper, V., et al. (1996). Degradation of cartilage aggrecan by collagenase-3 (MMP-13)L. F E B S Lett. 380, 17-20. Freije, J.M., Diez-Itza, I., Balbin, M., et al. (1994). Molecular cloning and expression of collagenase-3, a novel human matrix metalloproteinase produced by breast carcinomas. J. Biol. Chem. 269, 16766-16773. Galazka, G., Windsor, L.J., Birkedal-Hansen, H., et al. (1996). APMA (aminophenylmercuric acetate) activation of stromelysin-1 involves protein interactions in addition to those with cysteine-75 in the propeptide. Biochemistry 35, 11221-11227. Goldberg, G.I., Wilhelm, S.M., Kronberger, A., et al. (1986). Human fibroblast collagenase: Complete primary structure and homology to an oncogene transformation-induced rat protein. J. Biol. Chem. 261, 6600-6605. Gomis-Ruth, F.X., Gohlke, U., Betz, M., et al. (1996). The helping hand of collagenase3 (MMP-13): 2.7 A crystal structure of its C-terminal haemopexin-like domain. J. Mol. Biol. 264, 556-566. Hasty, K.A., Jeffrey, J.J., Hibbs, M.S., et al. (1987). The collagen substrate specificity of neutrophil collagenase. J. Biol. Chem. 262, 48-52. Hasty, K.A., Pourmotabedd, T.F., Goldberg, G.I., et al. (1990). Human neutrophil collagenasemA distinct gene product with homology to other matrix metalloproteinases. J. Biol. Chem. 265, 1421-1424. He, C.S., Wilhelm, S.M., Pentland, A.P., et al. (1989). Tissue cooperation in a proteolytic cascade activating human interstitial collagenase. Proc. Nat'l. Acad. Sci. USA 86, 2632-2636. Henriet, P., Rousseau, G.G., Eeckhout, Y. (1992). Cloning and sequencing of mouse collagenase cDNA--Divergence of mouse and rat collagenases from the other mammalian collagenases. F E B S Lett. 310, 175-178. Hirose, T., Patterson, C., Pourmotabbed, T., et al. (1993). Structure-function relationship of human neutrophil collagenase--Identification of regions responsible for substrate specificity and general proteinase activity. Proc. Natl. Acad. Sci. USA 90, 2569-2573. Hunt, L.T., Barker, W.C., and Chen, H.R. (1987). A domain structure common to hemopexin, vitronectin, interstitial collagenase and a collagenase homolog. Protein Seq. Data Anal. 1, 21-26.

40

JOHN J. JEFFREY

Jeffrey, J.J. (1986). The biological regulation of collagenase activity. In "Regulation of Matrix Accumulation" (R.P. Mecham, ed.), pp. 53-92, Academic Press, New York. Jeffrey, J.J. (1991). Collagen and collagenase--Pregnancy and parturition. Semin. Perinatol. 15, 118-126. Jeffrey, J.J., Welgus, H.G., Burgeson, R.E., et al. (1983). Studies on the activation energy and deuterium effect of human skin collagenase on homologous collagen substrates. J. Biol. Chem. 258, 11123-11127. Jenne, D. (1991). Homology of placental protein 11 and pea seed albumin 2 with vitronectin. Biochem. Biophys. Res. Commun. 176, 1000-1006. Jenne, D., and Stanley, K.K. (1987). Nucleotide sequence and organization of the human S-protein gene: Repeating peptide motifs in the pexin family and a model for their evolution. Biochemistry 26, 6735-6742. Knauper, V., Docherty, A.J., Smith, B., et al. (1997). Analysis of the contribution of the hinge region of human neutrophil collagenase (HNC, MMP-8) to stability and collagenolytic activity by alanine scanning mutagenesis. F E B S Lett. 405, 60-64. Knauper, V., Osthues, A., Declerck, Y.A., et al. (1993). Fragmentation of human polymorphonuclear leukocyte collagenase. Biochem. J. 291, 847-854. Knauper, V., Will, H., Lopez-Otin, C., et al. (1996). Cellular mechanisms for human procollagenase-3 (MMP-13) activation: Evidence that MT1-MMP (MMP-14) and gelatinase-a (MMP-2) are able to generate active enzyme. J. Biol. Chem. 271,17124-17131. Krane, S.M., Byrne, M.H., Lemaitre, V., et al. (1996). Different collagenase gene products have different roles in degradation of type-1 collagen. J. Biol. Chem. 271, 8509-8515. Lazarus, G.S., Brown, R.S., Daniels, J., et al. (1968). Human granulocyte collagenase. Science 159, 1483-1485. Li, J., Brick, P., O'Hare, M.C., et al. (1995). Structure of full-length porcine synovial collagenase reveals a C-terminal domain containing a calcium-linked, four-bladed bpropeller. Structure 3, 541-549. Liu, X., Wu, H., Byrne, M., et al. (1995). A targeted mutation at the known collagenase cleavage site in mouse type I collagen impairs tissue remodeling. J. Cell Biol. 130, 227-237. Lovejoy, B., Cleasby, A., Hassell, A.M., et al. (1994). Structure of the catalytic domain of fibroblast collagenase complexed with an inhibitor. Science 263, 375-377. McKerrow, J.H. (1987). Human fibroblast collagenase contains an amino acid sequence homologous to the zinc-binding site of serratia protease. J. Biol. Chem. 262, 5943. Miller, E.J., Finch, J.E., Chung, E., et al. (1976). Specific cleavage of the native type III collagen molecule with trypsin. Similarity of the cleavage products to collagenaseproduced fragments and primary structure at the cleavage site. Arch. Biochem. Biophys. 173, 631-637. Miller, E.J., Harris, E.D., Chung, E., et al. (1976). Cleavage of type I and II collagens with mammalian collagenase: Site of cleavage and primary structure at the NH2 terminal portion of the smaller fragment released from both collagens. Biochemistry 15, 787-792. Mitchell, P.G., Magna, H.A., Reeves, L.M., et al. (1996). Cloning, expression, and typeII collagenolytic activity of matrix metalloproteinase-13 from human osteoarthritic cartilage. J. Clin. Invest. 97, 761-768. Murphy, G., Allan, J.A., Willenbrock, F., et al. (1992). The role of the C-terminal domain in collagenase and stromelysin specificity. J. Biol. Chem. 267, 9612-9618. Murphy, G., Reynolds, J.J., Bretz, U., et al. (1977). Collagenase is a component of the specific granules of human neutrophil leukocytes. Biochem. J. 162, 195-197. Nagase, H. (1994). Matrix metalloproteinases: A mini-review. Extracell. Matrix Kidney Contrib. Nephrol. 85-93.

INTERSTITIAL COLLAGENASES

41

Ohuchi, E., Imai, K., Fujii, Y., et al. (1997). Membrane type-1 matrix metalloproteinase digests interstitial collagens and other extracellular matrix macromolecules. J. Biol. Chem. 272, 2446-2451. Quinn, C.O., Scott, D.K., Brinckerhoff, C.E., et al. (1990). Rat collagenase--Cloning, amino acid sequence comparison and parathyroid hormone regulation in osteoblastic cells. J. Biol. Chem. 265, 2342-2347. Reboul, P., Pelletier, J.P., Tardif, G., et al. (1996). The new collagenase, collagenase-3, is expressed and synthesized by human chondrocytes but not by synoviocytes. A role in osteoarthritis. J. Clin. Invest. 97, 2011-2019. Roswit, W.T., Halme, J., and Jeffrey, J.J. (1983). Purification and properties of rat uterine procollagenase. Arch. Biochem. Biophys. 225, 285-295. Saari, H., Suomalinin, K., Lindy, O., et al. (1990). Activation of latent human neutrophil collagenase by reactive oxygen species and serine proteases. Biochem. Biophys. Res. Comm. 171, 979-987. Sanchez-Lopez, R., Alexander, C.M., Behrendt, O., et al. (1993). Role of zinc-bindingencoded and hemopexin domain-encoded sequences in the substrate-specificity ofcollagenase and stromelysin-2 as revealed by chimeric proteins. J. Biol. Chem. 268, 72387247. Sang, Q.A., and Douglas, D.A. (1996). Computational sequence analysis of matrix metalloproteinases. J. Protein Chem. 15, 137-160. Seltzer, J.L., Welgus, H.G., Jeffrey, J.J., et al. (1976). The function of calcium ion in the action of mammalian collagenase. Arch. Biochem. Biophys. 173, 355-361. Stricklin, G.P., Bauer, E.A., Jeffrey, J.J., et al. (1977). Human skin collagenase: Isolation of precursor and active forms from both fibroblast and organ cultures. Biochemistry 16, 1607-1615. Stricklin, G.P., Eisen, A.Z., Bauer, E.A., et al. (1978). Human skin fibroblast collagenase: Chemical properties of precursor and active forms. Biochemistry 17, 2331-2337. Stricklin, G.P., Jeffrey, J.J., Roswit, W.T., et al. (1983). Human skin fibroblast procollagenase: Mechanisms for activation by organomercurials and trypsin. Biochemistry 22, 61-70. Suzuki, K., Enghild, J.J., Morodomi, T., et al. (1990). Mechanisms of activation of tissue procollagenase by matrix metalloproteinase 3 (stromelysin). Biochemistry 29, 1026110270. Test, S.T., Weiss, S.J. (1986). The generation and utilization of chlorinated oxidants by human neutrophils. Adv. Free Rad. Biol. Med. 2, 91-116. Tyree, B., Seltzer, J.L., Halme, J., et al. (1981). The stoichiometric activation of human skin fibroblast procollagenase by factors present in human skin and rat uterus. Arch. Biochem. Biophys. 208, 440-443. Vallee, B.L., and Auld, D.S. (1992). Active zinc binding sites of zinc metalloenzymes. Matrix metalloproteinases and inhibitors. Matrix (Suppl.) 1, 5-19. Van Wart, H.E., Birkedal-Hansen, H. (1990). The cysteine switch: A principle of regulation of metalloproteinase activity with potential applicability to the entire matrix metalloproteinase gene family. Proc. Natl. Acad. Sci. USA 87, 5578-5582. Wahl, L.M. (1977). Hormonal regulation of macrophage collagenase activity. Biochem. Biophys. Res. Commun. 74, 838-845. Welgus, H.G., Burgeson, R.E., Wootton, J.A.M., et al. (1985). Degradation of monomeric and fibrillar type III collagens by human skin collagenase--Kinetic constants using different animal substrates. J. Biol. Chem. 260, 1052-1059. Welgus, H.G., Campbell, E.J., Cury, J.D., Eisen, A.Z., Senior, R.M., Wilhelm, S.M., and Goldberg, G.I. (1990). Neutral metalloproteinases produced by human mononuclear phagocytesmenzyme profile, expression and regulation during development. J. Clin. Invest. 86, 1556-1564.

42

JOHN J. JEFFREY

Welgus, H.G., Grant, G.A., Sacchettini, J.C., et al. (1985). The gelatinolytic activity of rat uterus collagenase. J. Biol. Chem. 260, 3601-3606. Welgus, H.G., Jeffrey, J.J., Eisen, A.Z. (1981b). Human skin fibroblast collagenase: Assessment of activation energy and deuterium isotope effect with collagenous substrates. J. Biol. Chem. 256, 9516-9521. Welgus, H.G., Jeffrey, J.J., and Eisen, A.Z. (1981a). The collagen substrate specificity of human skin fibroblast collagenase. J. Biol. Chem. 256, 9511-9515. Welgus, H.G., Jeffrey, J.J., Stricklin, G.P., et al. (1980). Characteristics of the action of human skin fibroblast collagenase on fibrillar collagen. J. Biol. Chem. 255, 6806-6813. Welgus, H.G., Kobayashi, D.K., Jeffrey, J.J. (1983). The collagen substrate specificity of rat uterine collagenase. J. Biol. Chem. 258, 4162-4165. Wilson, C.L., and Matrisian, L.M. (1996). MatrilysinmAn epithelial matrix metalloproteinase with potentially novel functions (review). Int. J. Biol. Chem. 28, 123-136. Woessner, J.F. ( 1991). Matrix metalloproteinases and their inhibitors in connective tissue remodeling. F A S E B J. 5, 2145-2154. Zhang, Y.N., Dean, W.L., and Gray, R.D. (1997). Co-operative binding of Ca2 + to human interstitial collagenase, assessed by circular dichroism, fluorescence, and catalytic activity. J. Biol. Chem. 272, 1444-1447.

Stromelysins 1 and 2 Hideaki Nagase Department of Biochemistry and Molecular Biology, University of Kansas Medical Center, Kansas City, Kansas 66160

I. Introduction II. Structure and Substrate Specificity A. Structure B. Substrate Specificity III. Activation Mechanisms A. Activation by Nonproteolytic Agents B. Activation by Proteinases C. Involvement of the N-Terminal Phenylalanine for the Expression of Full Enzymic Activity D. Activation of Other proMMPs by MMP-3 IV. Inhibitors A. Synthetic Inhibitors B. TIMPs C. a2-Macroglobulins V. Regulation of Gene Expression A. Transcriptional Regulation B. Post-Transcriptional Regulation VI. Biological and Pathological Roles A. Biological Roles B. Pathological Roles VII. Conclusions and Future Prospects References

I.

INTRODUCTION

Stromelysins I and 2 are closely related metalloproteases that belong to the matrixin family (Woessner, 1991; Nagase, 1996). They degrade various components of the extracellular matrix, but not the triple helical regions of interstitial collagens. These properties distinguish them from collagenases, which share structural homology with stromelysins. A noncollagenolytic metalloprotease activity was first recognized in the extract of human articular cartilage by Sapolsky et al. (1974) and in the conditioned medium of rabbit synovial fibroblasts by Werb and Reynolds (1974). The enzyme was subsequently purified by Galloway et al. (1983) from the conditioned medium of rabbit bone explants and it was calledproteoglycase because it digested the core protein of cartilage proteoglycans at neutral pH. Later, the enzyme was renamed stromely43 Matrix MetaUoproteinases

Copyright 9 1998 by Academic Press. All rights of reproduction in any form reserved.

44

HIDEAKI NAGASE

sin, denoting a stromal cell-derived metalloproteinase that degrades extracellular matrix (Chin et al., 1985). On the other hand, Vater et al. (1983) reported collagenase activator of a doublet of 52 and 53 kDa from rabbit synovial fibroblasts and it was shown to be the precursor of a metalloproteinase. A similar activator, called collagenase activator protein was isolated in bovine articular cartilage (Treadwell et al., 1986). In 1985, Matrisian et al. isolated a cDNA clone whose message was induced in rat fibroblasts by epidermal growth factor or viral transformation. The protein, called transin, was expressed and shown to have metalloproteinase activity (Matrisian et al., 1986). In 1986, Okada et al. purified two isoforms of a metalloproteinase of 28 and 45 kDa from the medium of human rheumatoid synoviocytes that were stimulated with macrophage-conditioned medium. They refer to this metalloproteinase as matrix metalloproteinase 3 (MMP-3) to distinguish it from collagenase (MMP-1) and gelatinase (MMP-2), which are found in the same medium. By cDNA cloning of rabbit collagenase activator (Fini et al., 1987), human stromelysin (Whitham et al., 1986; Wilhelm et al., 1987), and human MMP-3 (Saus et al., 1988) and by immunochemical studies, it was shown that they are the same metalloendopeptidase. In 1988, Muller et al. identified a cDNA clone that is closely related (78% identical in amino acid sequence) to stromelysin from the human breast cancer cell cDNA library. They called it stromelysin 2. The original stromelysin, now called stromelysin 1, and stromelysin 2 are designated MMP-3 and MMP-10, respectively, following the numerical distinction of matrix metalloproteinases in the matrixin family (Nagase et al., 1992). An acid metalloproteinase purified from human cartilage was first designated as MMP6 (Azzo and Woessner, 1986), but it was proven to be MMP-3 (Wilhelm et al., 1993). This chapter reviews structure and function, gene regulation, and biological and pathological roles of stromelysins i and 2. The purification procedures, enzyme properties, and assays of human stromelysins 1 and 2 have already been reviewed (Nagase, 1995). II.

STRUCTURE

AND SUBSTRATE

A.

SPECIFICITY

Structure

Stromelysin 1 (MMP-3) and stromelysin 2 (MMP-10) are synthesized as pre-proenzymes and secreted from cells as proenzyme forms (proMMP-3 and proMMP-10). The primary structure of MMP-3 was deduced from cDNA clones of human (Whitham et al., 1986; Wilhelm et al., 1987; Saus et al., 1988), rabbit (Fini et al., 1987), rat (Matrisian

45

STROMELYSINS 1 AND 2

et al., 1985), and mouse (Ostrowski et al., 1988; Hammani et al., 1992), and those of MMP-10 are reported for human (Muller et al., 1988) and rat (Breathnach et al., 1987). Human proMMP-3 and proMMP-10 consist of a propeptide (82 amino acids; 79% identical), a catalytic domain (165 amino acids; 86% identical), a proline-rich "hinge region" (25 amino acids; 52% identical); and a C-terminal domain (188 amino acids; 75% identical) (see Fig. 1). The propeptide sequences contain the "cysteine switch" sequence PRCGVPD conserved in all matrixins, and the catalytic domains have the zinc-binding motif tIEXXHXXGXXH conserved among metzincin metalloproteinases (Bode et al., 1993; StScker et al., 1995). The C-terminal domains have a sequence that is

propeptide

I

Catalytic domain

]

I

Hinge region I I Hemopexin-like domain 45kDa

28kDa

I I

I

FIG. 1. Domain structures of prostromelysins 1 and 2 and the sites cleaved in the propeptide during activation. The box denotes the proteinase susceptible bait region. The conserved "cysteine switch" sequence PRCGVPD is boxed with a dashed line. The regions corresponding to a-helices are indicated (Becker et al., 1995). Cleavages induced by APMA treatment are shown by A (Nagase et al., 1990; Cameron et al., 1995). The final activation is mediated by the action of MMP-3 or MMP-10 intermediates generated by proteinases or APMA, and a 45-kDa stromelysin is generated. The 45-kDa form undergoes autolysis to the 28-kDa enzyme (Nagase et al., 1990; Suzuki et al., 1997). Trypsin (Tn) cleaves the ArgS4-Thr 85 bond and generates a 45-kDa species that exhibits about 20% of the activity (Benbow et al., 1996).

46

HIDEAKI NAGASE

similar to that of hemopexin and vitronectin. Stromelysin 3 (MMP-11) (Basset et al., 1990) diverges significantly from stromelysins 1 and 2 in amino acid sequence (Murphy, G. J. P., et al., 1991) and in enzymic activity. The crystal structure of the C-terminal domain-truncated proMMP3 has revealed that the pro-domain and the catalytic domain have separate folding units (Becker et al., 1995). The pro-domain consists of three a-helices and an extended peptide around the PRCGVPD (73-79) sequence. The overall folding of the catalytic domain of MMP-3 is very similar to interstitial collagenase (MMP-1) and neutrophil collagenase (MMP-8) (see StScker et al., 1995; Browner et al., 1995). It consists of five-stranded fl-sheet, three a-helices and connecting loops. It contains two molecules of zinc and at least two molecules of calcium (StScker et al., 1995). One zinc molecule is located at the active site of the enzyme interacting with side chains of His 2~ His 2~ and His 211, and the other is a structural zinc interacting with side chains of Asp 153,His 15~,His 1~, and His ~79 (Becker et al., 1995). Studies by Wetmore and Hardman (1996) propose that the second zinc is critical for maintaining the active form of MMP-3 by pulling the contiguous loops, which form the second zinc and first calcium-binding site (His 151 t h r o u g h Hisl~6), away from the active site and helping to organize the active site pocket, particularly the Sl' site, for substrate binding. About 2 mM Ca 2§ is required to stabilize proMMP-3 and active MMP-3 (Housley et al., 1993). The solution structure of the catalytic domain of MMP-3 has been also determined by nuclear magnetic resonance (NMR) imaging (Gooley et al., 1994; Van Doren et al., 1995). Biochemical and biophysical studies of proMMP activation suggest that the side chain of Cys in the conserved PRCGVPD motif interacts with the catalytic Zn 2§ as a fourth ligand to maintain the latency of proMMPs (Springman et al., 1990; Van Wart and Birkedal-Hansen, 1990; Salowe et al., 1992; Holz et al., 1992). The crystal structure of the C-terminal-truncated proMMP-3 (Becker et al., 1995) indicates that the region from Lys 72 to Va177 of KPRCGVPD occupies the active site cleft of the catalytic domain in a manner similar to that of a substrate forming identical fi-strand-like hydrogen bonds. However, the direction of this peptide is opposite from that of the substrate. The Cys 75 of this region interacts with the catalytic zinc as predicted by other studies. A similar three-dimensional structure can be predicted for proMMP-10 because they are closely related in the primary structure. The three-dimensional structures of the C-terminal domains of stromelysins have not been determined, but they are likely to be similar to those of collagenases (Li et al., 1995; Gomis-R~th et al., 1996) and gelatinase A (MMP-2) (Libson et al., 1995; Gohlke et al., 1996), which consist of four units of four-stranded antiparallel fl-sheet

STROMELYSINS 1 AND 2

47

stabilized on its fourfold axis by a cation (thought to be a calcium ion), forming a four-bladed/~-propeller structure. The presence of the Cterminal hemopexin-like domain is essential for collagenolytic activities of collagenases, but it does not influence the activity of MMP-3 on various substrates (Okada et al., 1986). ProMMP-3 secreted from human fibroblasts is partially glycosylated and a doublet of the 57-kDa (unglycosylated) and 59-kDa (glycosylated) forms is detected (Wilhelm et al., 1987; Okada et al., 1988). B.

Substrate

Specificity

MMP-3 digests various components of extracellular matrix. The information for MMP-10 is limited, but the enzyme also degrades a similar repertoire of matrix components although the catalytic efficiency is lower compared with stromelysin 1 (Nicholson et al., 1989; Murphy et al., 1991a; Nagase, 1995). The MMP-3 cleavage sites of various protein substrates are summarized in Table I. Other natural substrates include tenascin (Imai et al., 1994; Siri et al., 1995), vitronectin (Imai et al., 1995), perlecan (Whitelock et al., 1996), versican (Perides et al., 1995), laminin (Okada et al., 1986), elastin (Murphy et al., 1991a), and interleukin lfi (Ito et al., 1996). MMP-10 cleaves cartilage link protein at the His~-Ile 17and Leu25-Leu 2~ bonds (Nguyen et al., 1993). However, the activity of MMP-10 on fibronectin is negligible (Suzuki et al., 1997). MMP-3 exhibits an acid pH optimum activity around 5.5-6.0 for digestion of aggrecan and synthetic substrates, but it retains about 30-50% of the activity at pH 7.5 (Harrison et al., 1992; Wilhelm et al., 1993). MMP-10 has optimal activity against Azocoll and synthetic substrates at around pH 7.5-8.0 (Suzuki et al., 1997). In general, MMP-3 preferentially cleaves the peptide bond with a hydrophobic residue at the PI' position (Table I). However, the catalytic efficiency depends on the length of the peptide substrate. A peptide containing only three residues in the P site (the N-terminal side of the scissile bond) and two residues in the P' site (the C-terminal side of the scissile bond) is not readily cleaved unless the N-terminal and Cterminal ends are blocked (Table II) (Niedzwiecki et al., 1992). This suggests the enzyme has an extended substrate-binding site. MMP-3 accommodates substrates with aliphatic and aromatic residues at the P~' site, but MMP-1 fails to cleave substrates with aromatic residues at this site. This is due to the fact that MMP-3 has a larger and deeper S~' pocket than MMP-1. The residue at the P1 position is not involved significantly in contacting with the enzyme, but the catalytic efficiency increases when the P1 position has a charged group. The preferred residues at P3, P2 and P2' possessions are Pro, Leu, and aromatic resi-

TABLE I HUMAN STROMELYSIN 1 CLEAVAGESITES IN NATURAL SUBSTRATES Sequence Substrate

Source

P4

P3

P2

A g g r e c a n core p r o t e i n C a r t i l a g e link p r o t e i n Collagen a~ (II)

Human Human Bovine

Collagen O/1 ( I V ) Co llagen a2 (IV) Collagen ~1 ( I X ) Co llagen a2 (IX) Co llagen al (XI) Fibronectin a2-Macroglobulin

Bovine Bovine Bovine Bovine Bovine Human Human

Ovostatin al-Proteinase inhibitor l-Antichymotrypsin A n t i t h r o m b i n III ProMMP-1 ProMMP-3

Chicken Human Human Human Human Human

I R A Q (G (G (L (E Q P G R L E L I D D

P A G M P P A V A F P V N A L A V T

E I G G P I A A Q S E G A I S G A L

D

V

G

H s2

-

ProMMP-8 ProMMP-9

Human Human

D R

S V

G A

G 7s E 4~

-

D

L

G

R s7

-

P1

+

PI'

N 341

-

F 342

H 16 A 115 V 119

--

G1341) a

-

G143~ a

-

S59~)a $597)" A 482 p~s9 G ~79 F 6s4 G 677

_ -

p357

-

A 36~ R 393 QSO E 6s

-

117 Q1~6 M 12~ L 1342 F 1431 L 59s A 59s 1483 L 69~ L 6s~ y6s5 F 67s M 35s L 36~ S 394 F sl V 69 F s3 F 79 M 4~ F ss

_ -

-

--

P2'

P3'

P4'

Reference

F Q M Q K E K K L V R E T S V L V M R M R Q

G A G G G G R R Q A V S A I E N L R T L G T

V E V P L E P E Q T G D S P T P T K F T E F

F l a n n e r y et al., 1992 N g u y e n et al., 1989 W u et al., 1991 M o tt et al., 1997 M o t t et al., 1997 W u et al., 1991 W u et al., 1991 W u et al., 1991 W i l h e l m et al., 1993 E n g h i l d et al., 1989 E n g h i l d et al., 1989 M a s t et al., 1991 M a s t et al., 1991 M a s t et al., 1991 S u z u k i et al., 1990 N a g a s e et al., 1990 Kn~iuper et al., 1993 O g a t a et al., 1992

ProMMP-13

ProMMP-7 IGF-BP-3 Nidogen

Human Human Mouse

SPARC (BM-40, osteonectin)

Human

Fibulin-2

Mouse

a

H u m a n sequences.

L5S

Q6 A14 y16

F7 L15 L17

F Y V

G L C

L V G

A A A G F P A T R N A

E 77 y99 D82 N91 8351 p896 8976 T984 GLO38 81142 E21

y78 LlOO hS3 V92 y352 I897 L977 I985 11039 Vl143 V~2

S L D Y N N N I V I T

SO3 L L T Y T H H R T A E

F P T Y G Q G Q D N V

V S A S G

E 198 S209 p233 E543 G404

L199 L210 V234 M544 A405

L R

A V

O

O

E K

M

K L E

Substance P I n s u l i n B chain

Fibrin ~/-chain

G57

Human

Human

D L F G V Y G E N K T

V

O

R T K A A

Kmiuper et al., 1996a H a r r i s o n et al., 1989 Wilhelm et al., 1993

Imai et al., 1995 Fowlkes et al., 1994 M a y e r et al., 1993

Sasaki et al., 1997

Sasaki et al., 1996

Bini et al., 1996

TABLE HYDROLYSIS

OF SYNTHETIC

II PEPTIDES

BY MMP-3

Substrate

P6

P5

P4

P3

Dnp-

Pro

P2 -

P1

Tyr

-

Ala

PI' -

Tyr

P2' -

Tyr

P~' -

Met

P4' -

a

k cat

gm

k cat/Kin

( s -1)

(~M)

( s -1 M - ' )

Reference

Ps'

Arg

__

m

2,400 b

Netzel-Arnett

et al.,

1991 DnpMca

-

Pro

-

Lys

Pro

-

-

Pro

-Gln

Mca-

Leu-

Gly

-

Leu

-

Trp

-

Ala

-

D-Arg

-

NH2

--

m

2,200

Gln

-

Phe

-

Phe

-

Gly

-

Leu

-

Lys(Dnp)-Gly

0.53

50

10,900

-

Arg

-

NH2

Pro

-

Leu-

Gly

-

Leu

-

Dpa

-

Ala

Dnp

-

Arg

-

Pro

-

Lys

-

Pro

-

Leu-

Ala

-

Nva

-

Trp

-

NH2

Mca

-

Arg

-

Pro

-

Lys

-

Pro

-

Val

Glu

~

Nva

-

Trp

-

Arg

Mca

-

Arg

-

Pro

-

Arg

-

-

~

23,000

~

45,000 c

Lys(Dnp)-NH2

1.3 c

20 c

65,700 c

Lys

-

Pro

-Val

-

Glu

~

Nva

-

Trp

-

Arg

-

Lys(Dnp)-NH2

5.4

25

218,000

-

Pro

-

Phe

-

Glu

~

Leu

-

Arg

-

Ala

-

NH2

10

80

126,000

Pro

-

Lys

-

Pro

-

Gln

-Gln

~

Phe

-

Phe

-

Gly

-

Leu

-

Met-NH2

~

m

1790 c

Pro

-

Lys

-

Pro

-

Gln

-

-

Phe

-

Phe

-

Gly

-

Leu

-

Met-NH2

~

~

800 c

Lys

-

Pro

-

Gln

-Gln

-

Phe

-

Phe

-

Gly

-

Leu

-

Met-NH2

~

~

290 c

Pro

-

Gln-

Gln

-

Phe

-

Phe

-

Gly

-

Leu

-

Met-NH2

m

~